Multimedia Wireless Networks: Technologies, Standards, and QoS 9780130460998, 0130460990

Citation preview

< Day Day Up >

Table of Cont ent s

•

M u lt im e dia W ir e le ss N e t w or k s: Te ch n ologie s, St a n da r ds, a n d QoS By Aura Ganz, Zvi Ganz, Kit t i Wongt havarawat

Publisher: Prent ice Hall PTR Pub Dat e: Sept em ber 18, 2003 I SBN: 0- 13- 046099- 0 Pages: 352

From ent ert ainm ent t o t elephony, em erging wireless syst em s will m ake possible a new generat ion of wireless m ult im edia applicat ions. To sat isfy users, net work designers and developers m ust int egrat e end- t o- end Qualit y of Service ( QoS) support t hroughout all t heir underlying net works: WANs, WLANs, WPANs, and " last - m ile" WLL or sat ellit e dist ribut ion syst em s. However, wireless net work st andards t ypically focus on signaling, leaving crucial QoS issues t o im plem ent ers. Mult im edia Wireless Net works is t he first book t o help net work professionals syst em at ically address QoS in t oday's m ost im port ant wireless net works- and t om orrow's. ●

● ●

● ● ●

Why users' wireless m ult im edia perform ance requirem ent s will require ext ensive QoS support The fundam ent als of QoS - and how t hey drive net work design WLAN st andards from t he m ult im edia net work designer's viewpoint : I EEE 802.11, HiperLAN, and Hom eRF Wireless MANs: int roducing t he new 802.16 WirelessMAN st andard I nt egrat ing QoS int o I EEE 802.15 and Bluet oot h wireless personal area net works QoS in current and em erging cellular and sat ellit e net works < Day Day Up >

< Day Day Up >

Table of Cont ent s

•

M u lt im e dia W ir e le ss N e t w or k s: Te ch n ologie s, St a n da r ds, a n d QoS By Aura Ganz, Zvi Ganz, Kit t i Wongt havarawat

Publisher: Prent ice Hall PTR Pub Dat e: Sept em ber 18, 2003 I SBN: 0- 13- 046099- 0 Pages: 352

Copyright Prent ice Hall PTR Com m unicat ions Engineering and Em erging Technologies Series Preface Target Audience St ruct ure of t he Book Part 1. Mult im edia Applicat ions and Qualit y of Service ( QoS) Chapt er 1. Mult im edia Applicat ions Sect ion 1.1. Applicat ions Sect ion 1.2. Main Prot ocols Chapt er 2. Qualit y of Service Fundam ent als Sect ion 2.1. I nt roduct ion Sect ion 2.2. QoS Param et ers Sect ion 2.3. Mult im edia Applicat ion Requirem ent s Sect ion 2.4. QoS Services Sect ion 2.5. Realizat ion of QoS Services Chapt er 3. QoS Mechanism s Sect ion 3.1. I nt roduct ion Sect ion 3.2. Classificat ion Sect ion 3.3. Channel Access Mechanism Sect ion 3.4. Packet Scheduling Mechanism s Sect ion 3.5. Traffic Policing Mechanism Sect ion 3.6. Resource Reservat ion Signaling Mechanism s Sect ion 3.7. Adm ission Cont rol

Sect ion 3.8. QoS Archit ect ure Part 2. Wireless Local Area Net works Chapt er 4. I EEE 802.11 Sect ion 4.1. I EEE 802.11 Sect ion 4.2. I EEE 802.11e ( QoS Ext ension) Chapt er 5. HiperLAN Sect ion 5.1. I nt roduct ion Sect ion 5.2. Archit ect ure Sect ion 5.3. Physical Layer Sect ion 5.4. Dat a Link Cont rol ( DLC) Layer Sect ion 5.5. Convergence Layer Sect ion 5.6. QoS support Chapt er 6. Hom eRF Sect ion 6.1. I nt roduct ion Sect ion 6.2. Archit ect ure Sect ion 6.3. Physical Layer Sect ion 6.4. Media Access Cont rol ( MAC) Sect ion 6.5. QoS Support Part 3. Wireless Met ropolit an Area Net works Chapt er 7. I EEE 802.16 Sect ion 7.1. I nt roduct ion Sect ion 7.2. I EEE 802.16.1 Sect ion 7.3. Physical Layer Sect ion 7.4. Media Access Cont rol ( MAC) Sect ion 7.5. QoS Support Sect ion 7.6. I EEE 802.16a Part 4. Wireless Personal Area Net works Chapt er 8. Bluet oot h Sect ion 8.1. I nt roduct ion Sect ion 8.2. Archit ect ure Sect ion 8.3. Physical Layer Sect ion 8.4. Bluet oot h Baseband Sect ion 8.5. Link Manager ( LM) Sect ion 8.6. Host Cont rol I nt erface ( HCI ) Sect ion 8.7. Logical Link Cont rol and Adapt at ion Prot ocol ( L2CAP) Sect ion 8.8. Higher Bluet oot h Layers Sect ion 8.9. Profiles Sect ion 8.10. QoS Support Chapt er 9. I EEE 802.15 Sect ion 9.1. I EEE 802.15.1 Sect ion 9.2. I EEE 802.15.3

Sect ion 9.3. I EEE 802.15.3 Physical Layer Sect ion 9.4. I EEE 802.15.3 Media Access Cont rol Sect ion 9.5. I EEE 802.15.3 QoS Support Sect ion 9.6. I EEE 802.15.4 Sect ion 9.7. I EEE 802.15.4 Physical Layer Sect ion 9.8. I EEE 802.15.4 Media Access Cont rol Sect ion 9.9. I EEE 802.15.4 QoS Support Part 5. 2.5G and 3G Net works Chapt er 10. GPRS Sect ion 10.1. I nt roduct ion Sect ion 10.2. GPRS ( Rel- 5) Archit ect ure Sect ion 10.3. Physical Channel Sect ion 10.4. Logical, Cont rol, and Traffic Channels Sect ion 10.5. Media Access Cont rol ( MAC) and Radio Link Cont rol ( RLC) Sect ion 10.6. Radio Resource Cont rol ( RRC) and Radio Resource ( RR) Sect ion 10.7. QoS Support Chapt er 11. UMTS Sect ion 11.1. I nt roduct ion Sect ion 11.2. UMTS Archit ect ure Sect ion 11.3. Physical Layer Sect ion 11.4. Media Access Cont rol ( MAC) Sect ion 11.5. Dat a Link Layer Prot ocols ( RLC, PDCP, and BMC) Sect ion 11.6. Radio Resource Cont rol ( RRC) Sect ion 11.7. QoS Support Chapt er 12. cdm a2000 Sect ion 12.1. I nt roduct ion Sect ion 12.2. cdm a2000 Archit ect ure Sect ion 12.3. Physical Layer Sect ion 12.4. Media Access Cont rol ( MAC) Sect ion 12.5. Link Access Cont rol ( LAC) Sect ion 12.6. QoS Support Chapt er 13. Sat ellit e Com m unicat ion Sect ion 13.1. I nt roduct ion Sect ion 13.2. Archit ect ure Sect ion 13.3. Forward Link Sect ion 13.4. Ret urn Link Sect ion 13.5. Qualit y of Service Support Appendix Acronym s and Abbreviat ions Part 1: Mult im edia Applicat ion and Qualit y of Service ( QoS) Part 2: Wireless Local Area Net works Part 3: Wireless Met ropolit an Area Net works

Part 4: Wireless Personal Area Net works Part 5: 2.5G and 3G Net works Bibliography I nt roduct ion Part 1: Mult im edia Applicat ions and Qualit y of Service ( QoS) Part 2: Wireless Local Area Net works Part 3: Wireless Met ropolit an Area Net works Part 4: Wireless Personal Area Net works Part 5: 2.5G and 3G Net works < Day Day Up >

< Day Day Up >

Copyright Libr a r y of Con gr e ss Ca t a login g- in - Pu blica t ion D a t a A cat alog record for t his book can be obt ained from t he Library of Congress Publisher: Bernard Goodwin Edit orial/ product ion supervision: Nicholas Radhube Cover design direct or: Jerry Vot t a Cover design: Nina Scuderi Manufact uring m anager: Maura Zaldivar Edit orial assist ant : Michelle Vincent i Market ing m anager: Dan DePasquale © 2004 by Pearson Educat ion, I nc. Publishing as Prent ice Hall Professional Technical Reference One Lake St reet Upper Saddle River, New Jersey 07458 Prent ice Hall books are widely used by corporat ions and governm ent agencies for t raining, m arket ing, and resale. Th e pu blish e r offe r s e x ce lle n t discou n t s on t h is book w h e n or de r e d in qu a n t it y for bu lk pu r ch a se s or spe cia l sa le s. For m or e in for m a t ion , ple a se con t a ct : U.S. Cor por a t e a n d Gove r n m e n t Sa le s 1-800-382-3419 cor psa le s@pe a r son t e ch gr ou p.com For sa le s ou t side t h e U.S., ple a se con t a ct : I n t e r n a t ion a l Sa le s 1-317-581-3793 in t e r n a t ion a l@pe a r son t e ch gr ou p.com

Product and com pany nam es m ent ioned herein are t he t radem arks or regist ered t radem arks of t heir respect ive owners. All right s reserved. No part of t his book m ay be reproduced, in any form or by any m eans, wit hout perm ission in writ ing from t he publisher. Print ed in t he Unit ed St at es of Am erica First print ing. Pearson Pearson Pearson Pearson Pearson Pearson Pearson Pearson

Educat ion LTD. Educat ion Aust ralia PTY, Lim it ed Educat ion Singapore, Pt e. Lt d. Educat ion Nort h Asia Lt d. Educat ion Canada, Lt d. Educación de Mexico, S.A. de C.V. Educat ion—Japan Educat ion Malaysia, Pt e. Lt d.

Dedication This page is dedicat ed t o t he aut hors who slog t hrough t he t em plat e and use soft ret urns t o cent er t he t ext at logical, yet at t ract ive, break point s while picking up t he rules aut om at ically from t he reference page. < Day Day Up >

< Day Day Up >

Prentice Hall PTR Communications Engineering and Emerging Technologies Series Theodore S. Rappaport , Series Edit or D OSTERT Powerline Com m unicat ions D URGI N Space–Tim e Wireless Channels GANZ , GANZ , QoS

AND

W ONGTHAVARAWAT Mult im edia Wireless Net works: Technologies, St andards, and

GARG Wireless Net work Evolut ion: 2G t o 3G GARG I S- 95 CDMA and cdm a2000: Cellular/ PCS Syst em s I m plem ent at ion GARG & W I LKES Principles and Applicat ions of GSM

HA

Mult im edia Applicat ions Support for Wireless ATM Net works

KI M Handbook of CDMA Syst em Design, Engineering, and Opt im izat ion LI BERTI & RAPPAPORT Sm art Ant ennas for Wireless Com m unicat ions: I S- 95 and Third Generat ion CDMA Applicat ions PAHLAVAN & KRI SHNAMURTHY Principles of Wireless Net works: A Unified Approach RAPPAPORT Wireless Com m unicat ions: Principles and Pract ice, Second Edit ion RAZAVI RF Microelect ronics REED Soft ware Radio: A Modern Approach t o Radio Engineering STARR, CI OFFI & SI LVERMAN Underst anding Digit al Subscriber Line Technology STARR, SORBARA, CI OFFI & SI LVERMAN DSL Advances TRANTER, SHANMUGAN, RAPPAPORT & KOSBAR Principles of Com m unicat ion Syst em s Sim ulat ion wit h Wireless Applicat ions < Day Day Up >

< Day Day Up >

Preface The int roduct ion of wireless com m unicat ion is dram at ically changing our lives. The abilit y t o com m unicat e anyt im e, anywhere increases our qualit y of lives and im proves our business product ivit y. The recent t echnological developm ent s t hat allow us t o execut e bandwidt h- hungry m ult im edia applicat ions over t he wireless m edia add new dim ensions t o our abilit y t o com m unicat e. This opens an array of excit ing opport unit ies in business, resident ial, healt hcare, educat ion, leisure, and m any ot her areas. Wireless videoconferencing will connect us wit h business part ners and fam ily m em bers. Rem ot e video m edical consult at ion will enhance care in rural areas and at t he accident scene. I nt eract ive gam es t hat include video and graphics wit h part ners over t he globe will add new dim ensions not only t o our leisure opport unit ies but also t o provisioning of an effect ive rem ot e learning environm ent . Such opport unit ies are possible due t o t he recent t echnology developm ent s in 1) user device m iniat urizat ion, which enables adequat e com put at ion power and display in sm all m obile handheld devices, and 2) provisioning of significant ly broader wireless links for carrying m ult im edia t raffic. Such broader links have been int roduced in t he wide- ranging wireless net works including very short - range personal wireless net works, short - range local area net works, and longer range m et ropolit an, cellular, and sat ellit e net works. Because of nat ure of t he wireless environm ent , t hese links are shared am ong m any users execut ing m ult iple applicat ions, each requiring different levels of qualit y of service ( QoS) support . Therefore, each wireless net work needs t o incorporat e bandwidt h m ediat ion policies t hat enable QoS support t o t he different m ult im edia applicat ions. I t is int erest ing t o not e t hat t he wireless net works' st andards do not provide t he algorit hm s required for such bandwidt h m ediat ion policies. Such policies are t he m ot ivat ion for writ ing t his book. Wireless net works are described in num erous public dom ain docum ent s produced by st andard organizat ions in t he U.S., Europe, Asia, and ot her cont inent s whose m em bers include hundreds of part icipant s represent ing com panies all over t he world. These organizat ions include I EEE ( I nst it ut e of Elect rical and Elect ronics Engineers) , ETSI ( t he European Telecom m unicat ions St andards I nst it ut e) , and I TU ( I nt ernat ional Telecom m unicat ion Union) . The fact t hat t hese st andards are produced by such an im pressive collaborat ion of part icipant s and are public dom ain is t he basis for m aking t hese st andards a t rue, easy m eans for global com m unicat ion. Next generat ions of wireless net works include Wireless Personal Area Net works ( WPANs) , Wireless Local Area Net works ( WLANs) , Wireless Met ropolit an Area Net works ( WMANs) , and cellular and sat ellit e net works ( see Figure 1) .

Figu r e 1 . Th e Big Com m u n ica t ion Pu zzle

The provision of a qualit y experience for t he end- users requires end- t o- end QoS support in t erm s of bandwidt h, delay, and delay j it t er. I n order t o provide end- t o- end QoS support , we need t o provide QoS in t he wide area net work ( t he I nt ernet ) as well as in t he wireless ext ensions ( WPAN, WLAN, WMAN, and cellular and sat ellit e net works) . This book is a com prehensive guide t o underst anding m ult im edia wireless net works. The book addresses t he QoS problem s and solut ions, discussing t he archit ect ure, applicat ions, and im plem ent at ion of wireless net works, including a num ber of st andards and proposed st andards. I n t his book we int roduce t he basic QoS support m echanism s and t he st andards and st andardizat ion effort s in t he aforem ent ioned array of net work t echnologies. We focus on describing t he st andards' signaling m echanism s t hat need t o be incorporat ed in t he bandwidt h m ediat ion policies developed by net work designers. We hope t hat t he reader will realize t hat t he developm ent of algorit hm s t hat provide QoS support for different m ult im edia applicat ions is a very com plex t ask. We also hope t hat t he reader will underst and t he available signaling

m echanism s for each one of t he wireless st andards. Using t he knowledge base provided in t his book t he net work developers can t ake t he necessary st eps int o t he developm ent and im plem ent at ion of QoS m echanism s wit hin t heir t arget wireless net works. This book is not int ended t o replace t he st andards' docum ent s for any design purposes. I t s sole int ent ion is t o help and sim plify t he int roduct ion of such st andards and t o provide an easy t ool for m aking an init ial com parat ive analysis bet ween t he different approaches present ed by t he st andards in support ing m ult im edia t raffic. Since t he st andards cont ain hundreds and t housands of pages of det ailed inform at ion, using our personal j udgm ent , we had t o om it significant det ails, which we felt are less useful. Thus, we st rongly recom m end t hat for a det ailed analysis and design purposes, t he reader should refer t o t he official st andards' docum ent s. To t he best of our knowledge, t his is t he first book t hat covers such a broad array of wireless net works wit h a focus on m ult im edia support . We cert ainly hope t o cont inue and develop t he present ed t opics furt her by including m at erial t hat describes how net work designers and operat ors use t he t ools provided by t he st andards for providing users wit h an enj oyable m ult im edia experience. Hence, we will appreciat e your feedback and input via em ail t o Aura Ganz ( [email protected] ass.edu) . < Day Day Up >

< Day Day Up >

Target Audience The t arget audience of t his book includes net work m anagers, net work equipm ent developers, net work applicat ion developers, universit y st udent s, pract it ioners, and anyone who want s t o explore QoS issues in wireless net works. < Day Day Up >

< Day Day Up >

Structure of the Book This book is organized int o five prim ary part s.

Part 1: Multimedia Applications and Quality of Service (QoS) I n Chapt er 1 we explore t he nat ure of m ult im edia applicat ions t hat will be delivered t o t he users t hrough t he wireless net works. We discuss t he expect at ions t hat users have when ut ilizing t hese applicat ions. I n Chapt ers 2 and 3 we define QoS fundam ent al concept s and m echanism s and how t hey influence t he net work design. These QoS fundam ent als provide t he essent ial background required t o underst and t he QoS support aspect s of each wireless net work st andard in lat er chapt ers.

Part 2: Wireless Local Area Networks I n Part 2 we provide a closer look int o t he WLAN st andards and proposed st andards. Chapt ers 4, 5, and 6 cover I EEE 802.11, HiperLAN, and Hom eRF, respect ively.

Part 3: Wireless Metropolitan Area Networks I n Chapt er 7 we provide a closer look int o t he I EEE 802.16 WirelessMAN st andard.

Part 4: Wireless Personal Area Networks I n Part 4 we provide a closer look int o t he WPAN t echnologies, st andards, and proposed st andards. Chapt ers 8 and 9 describe Bluet oot h and I EEE 802.15, respect ively.

Part 5: 2.5G and 3G Networks I n Part 5 we provide a closer look int o t he 2.5G and 3G cellular net works as well as sat ellit e net work st andards and proposed st andards. Chapt ers 10, 11, and 12 cover General Packet Radio Service ( GPRS) , Universal Mobile Telecom m unicat ions Syst em ( UMTS) , and code division m ult iple access ( cdm a2000) , respect ively, and Chapt er 13 covers sat ellit e net works. < Day Day Up >

< Day Day Up >

Part 1: Multimedia Applications and Quality of Service (QoS)

< Day Day Up >

< Day Day Up >

Chapter 1. Multimedia Applications Sect ion 1.1. Applicat ions Sect ion 1.2. Main Prot ocols < Day Day Up >

< Day Day Up >

1.1 Applications 1.1.1 Evolution The landscape of I nt ernet usage is dram at ically changing because of t he rapid evolut ion of access t o wireless com m unicat ion wit h it s associat ed m obilit y. Consum ers in bot h business and resident ial m arket s are becom ing increasingly dependent on ubiquit ous access. Such access is enabled by t he proliferat ion of wireless com m unicat ion. Consum ers realize t he benefit s of m any new business applicat ions such as t hose in e- com m erce, collaborat ion, supply chain, and t elem edicine. They also st art t o appreciat e t he benefit s for t heir fam ily and personal needs such as com m unicat ion wit h friends, ent ert ainm ent , gam ing, and locat ion and safet y services. As t he speed and qualit y provisioning of wireless com m unicat ion increase, consum ers' dependency on applicat ions delivered t hrough t he wireless m edia will increase. They will also enj oy im proved graphics as t he wireless m edia will be able t o deliver cont ent - rich applicat ions t hrough t heir higher speed wireless net works. This new era of ubiquit ous wireless com m unicat ion will provide an array of applicat ions support ed by t he abilit y t o access t he I nt ernet everywhere and anyt im e. Wireless com m unicat ion will be available at hom e and at t he office. For exam ple, Microsoft inst alled a wireless local area net work ( WLAN) t hat let s workers connect t o t he corporat e int ranet from any spot on it s 265- acre Redm ond, Washingt on, cam pus. Such accessibilit y is becom ing possible also during t ravel via aut om obile, plane, ship, and t rain. Wireless connect ivit y support s Telem at ics, which allows drivers t o access t he I nt ernet via a screen in t heir cars. Whet her one walks in t he cit y, suburb, or rem ot e desert , wireless connect ivit y t o t he I nt ernet can be est ablished via sat ellit es provided by com panies such as Hughes Spaceway, as t hey can cover any point on t he cont inent . Ot her t ypes of global net works are envisioned once high- speed 3G cellular phones becom e available by com panies such as AT&T ( Unit ed St at es) , Vodafone ( Unit ed Kingdom ) , Orange ( France) , Om nit el ( I t aly) , and DoCoMo ( Japan) . I n t he regional dom ain, Wireless Local Loop t echnologies will provide t he basis for connect ion. WLANs provided by m any com panies such as 3Com , Linksys, and Cisco will cover local areas. Wireless Personal Area Net works ( WPANs) provided by com panies such as Ericsson will provide an opt ion for coverage of personal areas in which com m unicat ion needs t o be est ablished bet ween neighboring devices. Som e of t he applicat ions t hat current ly use wireless net works as t heir t ransm ission m edia were previously available via wired m edia and som e applicat ions are new. Here is a part ial list of t hese applicat ions: ● ● ● ● ● ● ● ● ● ● ● ●

●

St ream ing video St ream ing audio Collaborat ion One- way and int eract ive m ult im edia m essaging Gam ing, including int eract ive peer- t o- peer ( p2p) gam ing Digit al m oney t ransact ions MP3 m usic download Video- and audio- support ed shopping Long- dist ance learning, educat ion Video and audio conferencing File sharing and t ransfer ( pict ures, video clips, and t ext ) Feeding of real- t im e news and inform at ion about t he weat her, financial m arket s, sport s and so on Geographic locat ion services

● ● ●

Safet y services such as Enhanced 911 ( E911) Gam bling Ent ert ainm ent

These applicat ions will be delivered t o consum ers in m any shapes and form s depending on t he devices and wireless com m unicat ion m edia available t o t he consum er. The wireless net work has a profound role in t he effect iveness of t he applicat ion delivery. When higher speed connect ions are available, t he applicat ions can deliver richer cont ent , im proved graphics, and m ore vivid colors. When t he net work can support Qualit y of Service ( QoS) , t he applicat ions can im prove int eract ivit y, reduce j it t er, and provide cont inuous video and voice experience. When lat ency is high, applicat ions t hat t ailor locat ion- based inform at ion for consum ers m ay provide out dat ed inform at ion. Likewise, t he device archit ect ure has a st rong effect on t he consum er experience. For exam ple, a larger screen can accom m odat e larger pict ures and m ore inform at ion. This dim ension is am plified in t he wireless environm ent , since devices t end t o have m uch sm aller screens. A m ore powerful cent ral processing unit ( CPU) can expedit e com put at ions associat ed wit h t he applicat ions, it s graphics, and various com m unicat ion prot ocols. More m em ory can enhance t he graphic visualizat ion experience. Moreover, such m em ory can be used for caching various m ult im edia cont ent s for fut ure view. The pot ent ial list of applicat ions t hat can benefit consum ers in t he wireless m obile environm ent is long and covers all aspect s of life and business. We describe som e applicat ions in order t o give an idea of t he broad range of benefit s t hat consum ers can expect .

1.1.2 Video and Audio Streaming Video and audio st ream ing provides t he m eans of delivering news, ent ert ainm ent , rem ot e educat ion, docum ent ary, corporat e speeches, fashion shows, and m any m ore t ypes of com m unicat ion. Television m ay be t he m ost well- known form of st ream ing video. I t already feeds wireless m ult im edia st ream s int o m illions of dishes and ant ennas, connect ed t o TVs and ot her devices. DirecTV ( www.direct v.com ) and Dish Net works ( www.dishnet works.com ) are t wo m aj or providers of st ream ing video in t he Unit ed St at es. St ream ing t echnologies are im port ant , especially in t he wireless world, since m ost users do not have access t o enough connect ion capacit y t o download large m ult im edia files quickly. Using st ream ing t echnologies, consum ers can st art list ening t o t he audio st ream or view t he video st ream before t he ent ire file has been received. To allow efficient st ream ing, t he provider needs t o send t he dat a as a st eady st ream and t he receiver needs t o be able t o cache excess dat a in a t em porary buffer unt il used. I f t he dat a do not arrive fast enough, users will experience int errupt ions. There are several com pet ing st ream ing t echnologies, such as RealAudio, RealVideo ( www.real.com ) , Microsoft Media Player ( www.m icrosoft .com ) , Packet Video ( www.packet video. com ) , and QuickTim e ( www.apple.com ) . To reduce t he am ount of inform at ion t ransm it t ed, st ream ing video and audio dat a are com pressed by m eans of t echnologies such as MPEG. The st ream ing video qualit y depends on t he capacit y of t he t ransm ission channel and it s abilit y t o support a st eady st ream —t he bet t er t he channel qualit y ( i.e., higher and st eady dat a rat e) , t he bet t er t he qualit y of t he audio and video out put .

1.1.3 Peer-to-Peer Computing

Peer- t o- peer ( p2p) com put ing is t he sharing of files, m em ory, com put at ion power, and ot her com put er resources and services am ong devices. p2p Com put ing aggregat es t he shared net work resources allowing econom ical execut ion of applicat ions. For exam ple, inst ead of purchasing addit ional m em ory for a PDA, t he user can benefit from available m em ory on ot her devices. I n a p2p archit ect ure devices can com m unicat e direct ly wit h t heir peers wit hout t he need for a cent ral server. They can support collaborat ion am ong users on different wireless and wired net works, m oving dat a closer t o t he end user. This collaborat ion is obt ained via caching m echanism s and dist ribut ed com put ing using t he com bined power of CPUs and m em ory of t he devices t hat part icipat e in t he p2p net work. p2p Com put ing has an im port ant role in bot h business and resident ial applicat ions. Napst er st art ed t he revolut ion in m usic dist ribut ion. Com panies such as Groove, Endeavors, and eZm eet ing are developing p2p business applicat ions. Such applicat ions support collaborat ion wit h rich cont ent . Using p2p collaborat ion soft ware, consum ers can est ablish p2p work groups ( see Figure 1.1) t hat are accessible only by predet erm ined or invit ed m em bers. Through t he click of a but t on, workgroup m em bers can browse t he web t oget her, share docum ent s, edit pict ures, t alk t o each ot her t hrough a VoI P ( voice over I nt ernet prot ocol) feat ure, share calendars, or draw diagram s on a dedicat ed whit e board.

Figu r e 1 .1 . Pe e r - t o- Pe e r Colla bor a t ion

1.1.4 Digital Money Transactions Digit al m oney t ransact ions involving wireless devices and net works have spurred a variet y of applicat ions.

Banks will be able t o support t he consum er who want s t o obt ain cash from a t radit ional aut om at ed t eller m achine ( ATM) using his or her m obile phone. To save t im e, t he consum er is able t o st art t he wit hdrawal t ransact ion process prior t o approaching t he ATM. The consum er can input t he account num ber and choose t he t ransact ion t ype and am ount t o be wit hdrawn. When t he consum er reaches t he ATM, he or she ent ers t he securit y personal ident ificat ion num ber int o t he m obile phone and t he t ransact ion det ails are t ransm it t ed t o t he ATM. The ATM processes t he t ransact ion and provides t he am ount of cash ent ered. Spar Nord Bank and Laan & Spar Bank in Denm ark are working on such t echnologies. DoCoMo and com panies such as Coca- Cola int roduce Cm ode as part of t heir i- m ode service. Cm ode is a shopping syst em t hat uses digit al t ransact ions and t hat links i- m ode phones t o Cm ode vending m achines. Consum ers can use t heir m obile phones t o purchase product s from Cm odecom pliant m achines such as vending m achines which include t heir own com put er, display screen, speakers, print ers, and various sensors. To use t he Cm ode service, consum ers need t o regist er and deposit m oney int o a special account m aint ained by DoCoMo. Aft er t he consum er m akes a purchase, he or she receives a " C Ticket ," which shows up on t he phone's screen as a kind of bar code. When t his code is passed in front of t he vending m achine's sensor, t he m achine releases t he purchased product . Ot her com m unicat ion besides t he bar code t ransm ission can be envisioned. For exam ple, com m unicat ion via a Wireless Personal Net work t hat is est ablished bet ween t he i- m ode phone and t he vending m achine. I n Japan, Lawson ( one of t he largest chains of convenience st ores wit h m ore t han 7,500 shops) equipped t heir st ores wit h DoCoMo iConvenience using i- m ode service. Wit h iConvenience, shoppers can m ake purchases t hrough t heir cell phone. The shopper can ent er t he purchase num ber int o t he kiosk and t hen t ake t he print ed t icket and bar code t o t he cash regist er for paym ent .

1.1.5 Entertainment One of t he m ost prevalent applicat ions for WLANs is ent ert ainm ent . For exam ple, hom e t heat er syst em s dist ribut e m ovies t o t he m onit ors and speakers using wireless int erconnect ion. The video sources ( i.e., cable t elevision, DVD player) can be in one room but t he m onit or can be in anot her room . A wireless solut ion m akes it convenient t o set up t he net work so people do not need t o deal wit h wires. Digit al audio files ( i.e., MP3, cable t elevision, RealAudio) have recent ly becom e popular. I nst ead of buying m usic CDs, people can purchase, download or st ream digit al m usic files from t he I nt ernet t o t heir com put ers. So t hat t he use of t he ent ert ainm ent is not lim it ed t o only one com put er, som e product s provide a solut ion t o dist ribut e such digit al m usic t o t he hom e st ereo t hrough a wireless audio receiver ( Figure 1.2) .

Figu r e 1 .2 . W ir e le ss H om e En t e r t a in m e n t

1.1.6 File and Picture Sharing File sharing is a very com m on applicat ion in resident ial and corporat e set t ings. Com put ers in t he sam e work group or in t he sam e dom ain can access and share files st ored in each com put er. Present ly, file sharing is lim it ed not only t o a local area net work, but is also ext ended t o t he I nt ernet . There are t wo approaches for file sharing. One is cent ralized file sharing, in which t here are cent ral file servers operat ing as rem ot e hard drives or online st orage. A user uploads t he files t o a cent ral file server. Then ot her users are allowed t o access and download t he files from t he server via t he I nt ernet . For securit y and privacy reasons, before accessing t he files, users are required t o provide som e form of aut hent icat ion ( i.e., usernam e or password) . Com panies such as Xdrive ( www.xdrive.com ) , Yahoo! Briefcase ( www.briefcase.yahoo.com ) , and St oragePoint ( www. st oragepoint .net ) offer such online st orage services. The ot her t ype of file sharing is dist ribut ed file sharing, in which t here is no cent ral file server. The com put ers connect ing t o t he I nt ernet can share files in a p2p fashion ( as described in Sect ion 1.1.3) . A com put er can access and download files direct ly from anot her com put er on t he I nt ernet . Soft ware such as Kazaa Media Deskt op ( www.kazaa.com ) and Morpheus ( www.m orpheus.com ) enable t his kind of service. Digit al pict ures generat ed from t he digit al cam era or pict ure scanner also st im ulat e growt h of pict ure- sharing services. People want t o share t heir digit al pict ures wit h friends and fam ily. One sim ple solut ion is t o append t he pict ure as an em ail at t achm ent . However, such em ail files becom e very large, which leads t o long download t im es, especially via m odem . Com panies such as Clubphot o ( www.clubphot o.com ) , ofot o ( www.ofot o.com ) , and Yahoo! Phot os ( www.phot os. yahoo.com ) offer users t he opport unit y t o st ore and share t heir pict ures over t he I nt ernet . Som e of t hese services also include ext ra feat ures such as phot o album s, slide shows, and phot o print ing. These services sim plify t he pict ure- sharing process, since users need t o send t heir friends only t he URLs of t he rem ot e sit es.

1.1.7 Email and Multimedia Messaging The form at ion of com m unit ies has been a st rong desire of m any consum ers, especially t eens. These com m unit ies are form ed around enhanced m essaging services t hat allow part icipant s t o send m essages, pict ures, graphics, and video and audio files t o each ot her. Prot ocols such as Short Message Service ( SMS) , Mult im edia Messaging Service ( MMS) , and Enhanced Messaging Service ( EMS) define t he archit ect ure for various degrees of form at t ing, graphics, and pict ures t hat can be exchanged am ong part icipant s.

Com panies such as Spot Life ( www.spot life.com ) develop video com m unit ies for wireless carriers. Part icipant s will be able t o view and send personal video cont ent t o ot her part icipant s. This applicat ion can also be used for t he business environm ent . Cell phones equipped wit h cam eras will be able t o capt ure live video im ages.

1.1.8 Wireless Gaming The wide success of hand- held gam ing devices such as Gam eBoy and Cybiko has dem onst rat ed t he vit alit y of gam es being played on sm all devices in addit ion t o large- screen t elevisions and personal com put ers. Com panies such as Sega, Nint endo, and Disney have st art ed t o t arget m obile wireless devices as an im port ant part of t heir gam e developm ent st rat egy. Som e success st ories dem onst rat e t hat wireless device users have an appet it e for wireless m ult iplayer gam es. I n 2001 JAMDAT Mobile ( www.j am dat m obile.com ) report ed t hat 300,000 m obile device users spent m ore t han t hree m illion air m inut es in less t han t hree m ont hs wit h t heir m ult iplayer gam e Gladiat or. Gladiat or is a wireless m ult iplayer com bat gam e set in t he Coliseum of ancient Rom e. Players choose charact ers and weapons for t heir gladiat ors and t hen duel against a live opponent in real t im e. The m ore a player ut ilizes a part icular gladiat or, t he m ore experience and skill t his gladiat or obt ains. This com pany report ed t hat players accessed t heir propriet ary gam e server from m ore t han 20 wireless carriers around t he world using m ore t han 65 different devices. Different players who part icipat e in a m ult iplayer gam e m ay connect t o different wireless net works wit h different capabilit ies. Their server allowed users t o overcom e such incom pat ibilit ies, providing a sm oot h gam ing experience. There are m any ot her gam es t hat can be played over wireless net works, such as nGam e's Rat Race ( www.ngam e.com ) which allows a t eam of players t o " race" against one anot her using t he phone, Disney's At lant is: The Lost Em pire in which players creat e t eam s and search for t he Lost Cont inent by sending bubbles up from t he dept hs t o get t hree black or t hree whit e bubbles in a row horizont ally or vert ically, and Disney/ Pixar's Monst ers I nc. in which players cat ch scream s t o achieve st at us as a Monst ers I nc. professional kid scarer. Realizing t he opport unit y, Ericsson, Mot orola, Nokia, Siem ens, and ot her com panies founded t he Mobile Gam es I nt eroperabilit y Forum ( MGI F) ( www.m gif.org) . The forum 's goal is t o define specificat ions for m obile gam es developers so t hat t heir gam es can be played over various m obile devices and net works.

1.1.9 Voice and Telephony Wireless net works t hat provide voice and t elephony services include cellular phones, cordless phones and VoI P. I n 2002 t he num ber of cellular phone subscribers reached one billion worldwide. The cellular phone offers not only voice service but also dat a services in t he current 2.5G syst em and t he upcom ing 3G syst em . These syst em s allow users t o surf t he net or check em ail using t heir cellular phones. VoI P provides voice or t elephone service over a dat a net work ( e.g., wired LAN, I EEE 802.11 wireless LAN, and I nt ernet ) . As shown in Figure 1.3, users can t alk wit h each ot her via t he com put ers connect ed t o t he dat a net work. VoI P can be est ablished eit her bet ween com put ers or bet ween com put ers and regular phones. An exam ple of com put er- t o- com put er VoI P soft ware is MSN Messenger. Com panies such as Net 2Phone ( www.net 2phone.com ) and Dialpad ( www.dialpad. com ) offer com put er- t o- phone VoI P soft ware. An ent erprise can deploy VoI P t o reduce t he cost of long- dist ance calls bet ween branches ( see Figure 1.4) .

Figu r e 1 .3 . VoI P Sce n a r io

Figu r e 1 .4 . VoI P Sce n a r io for En t e r pr ise s

1.1.10 Location-Based Services Locat ion- based services are envisioned in m any applicat ions ranging from m ilit ary t o civilian everyday life. They include applicat ions such as t raffic report ing, rest aurant recom m endat ion, navigat ion, and cust om ized ads. I n such applicat ions t he inform at ion is sent t o a user based on t he user's locat ion provided via t he GPS ( Global Posit ioning Syst em ) . Unit ed St at es m ilit ary forces used it effect ively in Operat ion Desert St orm . Support ed by GPS, t hese forces were able t o navigat e t hrough t he wide, feat ureless desert in which t he t errain looks t he sam e for as m uch as t he eyes can see. GPS allowed t hese forces t o m aneuver t hrough sandst orm s and during night t im e. This navigat ion syst em is also used for navigat ion by com m ercial applicat ions. The const ruct ion of t he t unnel under t he English Channel was carried out sim ult aneously by Brit ish and French t eam s digging from opposit e ends of t he t unnel—t he English t eam from Dover, England, and t he French t eam from Calais, France. These t eam s relied on GPS receivers out side t he t unnel t o validat e t heir posit ions along t he way and t o ensure t hat t hey would m eet exact ly in t he m iddle of t he new t unnel. Sim ilarly, GPS locat ion inform at ion is being used in ships, t rucks, and cars. By m eans of video display on t he dashboard, t he vehicle locat ion dat a can be applied for navigat ion and cont rol. I n t he field of wildlife m anagem ent , endangered species are being t racked by GPS receivers t o m onit or, st udy, and cont rol populat ion. Locat ion inform at ion can also save hum an lives. One exam ple is E911 services. They enable callers t o use t heir cellular phones t o easily place 911 em ergency calls and provide t he em ergency services wit h t heir geographic locat ion. I n sit uat ions when t he caller is t raveling on an unfam iliar road, in an area wit h few landm arks, or at night , calling 911 on a m obile phone is probably t he only choice. Using E911 services, t he em ergency

t eam will have t he driver's exact locat ion, significant ly short ening t he t im e from when t he rescue t eam s are sent unt il t hey arrive at t he caller's locat ion. New- locat ion based applicat ions are invent ed as wireless devices equipped wit h GPS cont inue t o evolve. I m agine receiving in your cell phone ret ail coupons based on your locat ion. When a driver is passing a Burger King, t he cell phone will beep and display an ad t hat suggest s st opping in for one dollar off t hat new ham burger. The consum er can press keys t o accept or refuse t he offer or have an em ail sent for m ore inform at ion or fut ure redem pt ion. Com panies such as DoCoMo ( iarea service) , Vindigo, and Avant Go developed such services for handheld devices. These com panies report t hat m illions of users are regist ered for t hese services. These excit ing GPS- based locat ion services are enabled by a syst em of 24 sat ellit es which are const ant ly orbit ing t he eart h. They m ake t wo com plet e orbit s every 24 hours at approxim at ely 12,000 m iles above t he eart h's surface. The GPS sat ellit es cont inuously t ransm it digit al radio signals t hat cont ain dat a on t heir locat ions and t he exact t im e t o t he eart h- based receivers. These sat ellit es are equipped wit h at om ic clocks t hat are very precise ( less t han a billiont h of a second error) . The receivers on eart h know how long it t akes for t he signal t o arrive. Since t he signal t ravels at t he speed of light , t he receiver knows how far away t he sat ellit e is locat ed. By using dat a from t hree sat ellit es, t he receiver com put es it s locat ion in t erm s of longit ude and lat it ude. By receiving a signal from one m ore sat ellit e, m eaning four sat ellit es in t ot al, t he GPS receiver can also det erm ine it s alt it ude. The GPS syst em is based upon a direct line of sight bet ween t he receiver ( e.g., cell phone) and t he sat ellit es. That m eans GPS services are not working in big cit ies wit h t all buildings or in buildings. The FGDC ( Federal Geographic Dat a Com m it t ee) , which coordinat es t he developm ent of t he U.S. Nat ional Spat ial Dat a I nfrast ruct ure, has proposed a st andard for locat ion services. This st andard's obj ect ive is t o increase t he int eroperabilit y of locat ion services appliances wit h print ed m aps. The com m it t ee proposes t o est ablish a nat ionally consist ent grid reference syst em as t he preferred grid for Nat ional Spat ial Dat a I nfrast ruct ure applicat ions. This U.S. Nat ional Grid is based on universally defined coordinat es and grid syst em s and is pot ent ially ext endable as a universal grid reference syst em for worldwide applicat ions.

1.1.11 Telemedicine Wireless com m unicat ion has brought new opport unit ies t o save lives. For exam ple, in t he rem ot e part s of Canada, a m ult im edia sat ellit e proj ect , t he Rem ot e Com m unit ies Services Telecent re ( RCST) proj ect described in www.rcst .net , is run by a sat ellit e operat or, Telesat . RCST has been operat ing since 1998 linking a num ber of rural t elecent ers in Labrador and Newfoundland wit h ot hers in m ore populat ed areas. These t elecent ers provide int egrat ed t ele- learning and t elem edicine services via broadband sat ellit e com m unicat ion links. These services include highspeed I nt ernet access, video conferencing, and digit al im aging. A physician using t he RCST syst em in t he rem ot e Port aux Basques on t he west ern edge of Newfoundland can, for exam ple, t ransm it im ages of a pat ient 's condit ion t o a specialist in St . John's locat ed m ore t han 450 kilom et ers away. This specialist can help in diagnosing t he condit ion, recom m end t reat m ent , or suggest a physical visit . This syst em saves crit ical t im e as com pared t o t he t im e required for on- sit e diagnosis at t he specialist 's locat ion, which m ay be especially long in difficult weat her condit ions. I n addit ion, t he I nt egrat ed Em ergency Medicine Net work ( I EMN) syst em has been int roduced. This syst em provides em ergency m edical care for pat ient s during am bulance t ransit . These special am bulances allow physicians t o m onit or t he vit al signs of pat ient s in t ransit and bet t er prepare for t heir arrival at t he hospit al ( see Figure 1.5) .

Figu r e 1 .5 . Te le m e dicin e Sce n a r io

Sim ilar proj ect s have been im plem ent ed in Africa, t he Philippines, and locat ions where t he dispersed populat ion faces difficult long- dist ance com m ut es t o reach healt h care providers.

1.1.12 Business Applications Business applicat ions such as paym ent aut horizat ion, dist ance learning, m ult icast delivery of prom ot ional cont ent , and ret ail point - of- sale ( POS) t ransact ion processing can be handled cent rally via one sat ellit e com m unicat ion provider. For exam ple, General Mot ors and Ford are able t o t rain it s m anagem ent and t echnical personnel in any office, anywhere in t he world. Each st udent can inst ant ly access int eract ive dist ance learning m odules via DirecWay Hughes Sat ellit es. WalMart , Shell, Texaco, and ot her com panies use t his service for t heir ret ail POS t ransact ions.

1.1.13 Personal Area Device Connectivity Personal area devices refer t o devices carried by, placed near, or used near a person. There is a wide range of such devices: ●

Consum er devices ( Figure 1.6) : port able CD player, digit al cam era, print er

Figu r e 1 .6 . Con su m e r D e vice s

● ● ● ● ●

Gam es: gam e cont roller, j oyst ick Professional devices: personal digit al assist ant ( PDA) , not ebook, pager, cellular phone Sport t raining devices: healt h m onit or, sensor, m ot ion- t racking devices Hospit al devices: blood pressure sensor, heart rat e sensor, elect rocardiogram Milit ary devices: com bat equipm ent ( Figure 1.7)

Figu r e 1 .7 . Com ba t Equ ipm e n t

Typically, personal devices have propriet ary cables t hat int erconnect ( e.g., earphone, print er cable) . Som et im es t he cable is bulky, easily dam aged, and frust rat ing t o use. Wireless t echnologies becom e alt ernat ive solut ions t o replace t hese cables.

1.1.14 Telematics (Automobiles) New cars are designed wit h m ore and m ore digit al devices for every aspect including safet y, operat ion, com m unicat ion, and ent ert ainm ent . There are new digit al devices t hat cont rol every funct ion of t he car ranging from fuel inj ect ion t o airbag operat ion. I n addit ion, ent ert ainm ent opt ions have increased t o include sat ellit e radio, video, and gam ing. Wireless com m unicat ion, com bined wit h locat ion services provided by GPS, has enabled safet y and securit y services such as rout ing inst ruct ions, em ergency door unlocking, em ergency response, and st olen vehicle not ificat ion ( see Figure 1.8) .

Figu r e 1 .8 . Te le m a t ics Sce n a r io

Mercedes- Benz offers voice- act ivat ed phones, aut om at ic headlight s, and sensors t hat wat ch for nearby obst acles. BMW offers em ail access as well as it s iDrive syst em , which allows drivers hundreds of opt ions t o access t he car devices wit h a sm all cont rol knob and on- screen m enus. I n addit ion, BMW's Mayday syst em offers wireless com m unicat ions and GPS locat ion services for 24hour em ergency response, roadside assist ance, st olen vehicle not ificat ion, and rem ot e door unlocking. Sim ilar services are offered by ot her com panies such as General Mot ors. Their OnSt ar Syst em int egrat es on- board advanced vehicle elect ronics wit h GPS t echnology and wireless com m unicat ions t echnology. The OnSt ar Syst em operat es 24 hours a day and links t he driver and car t o a 24- hour response cent er for safet y, securit y, and convenience needs. Mot orola's iRadio proj ect dem onst rat es wireless delivery of a wide variet y of services. The iRadio converges int o one package t hat includes pict ure ent ert ainm ent , inform at ion, navigat ion, and com m unicat ion. Drivers and passengers can configure t heir personal preferences. I t s navigat ion syst em can provide real- t im e rout e planning and t ravel inst ruct ions, off- rout e not ificat ion, dest inat ion ident ificat ion, t raffic advisories, and local inform at ion. Radio is provided by t he sat ellit e radio channels. Users have access t o news, st ock quot es and ot her financial updat es. Telephony services including em ail access are also support ed. I nt ernet connect ion can be provided t hrough sat ellit es or cellular net work connect ions. Ford, Daim ler- Chrysler, General Mot ors, Renault , and Toyot a form ed t he Aut om ot ive Mult im edia I nt erface Collaborat ion ( AMI C) forum t o facilit at e t he developm ent , prom ot ion, and st andardizat ion of aut om ot ive m ult im edia int erfaces t o m ot or vehicle com m unicat ion net works. The goal is t o have an open st andard t hat will elim inat e com plex and incom pat ible syst em s t hat can only operat e on a single t ype of vehicle. The open st andard will m ake it possible for elect ronic devices t o be aut om at ically configured and t o com m unicat e wit h each ot her. The forum considers a variet y of wired and wireless t echnologies for t he net worked vehicle. < Day Day Up >

< Day Day Up >

1.2 Main Protocols 1.2.1 Short Message Service (SMS) Short Message Service ( SMS) is a t ext m essage service t hat enables t he t ransm ission of short m essages bet ween a cell phone, PDA, PC, or any device wit h an I nt ernet prot ocol ( I P) address. Since t his service was st andardized in t he m id 1990s, it has recorded t ens of billions of m essages and is being used worldwide. SMS was int roduced in t he Global Syst em for Mobile Com m unicat ions ( GSM) and lat er adopt ed by ot her m obile com m unicat ion syst em s such as Tim e Division Mult iple Access ( TDMA) and Code Division Mult iple Access ( CDMA) . The service support s m essages t hat are no m ore t han 140 t o 160 alphanum eric charact ers in lengt h, which is t ypically around 30 t o 40 words in Lat in and 70 charact ers for non- Lat in alphabet s like Chinese and Arabic. I nit ial applicat ions of SMS were based on replacing alphanum eric pagers by support ing t wo- way m essaging services, prim arily voice m ail. SMS is used for m any applicat ions. For exam ple, Gam eWorld Technologies' gam e ent it led " Wom en Are Sm art er?" is a m obile t rivia gam e wit h a gender t wist . Players respond t o t rivia quest ions t hat include t he opt ion t o m ake double- or- not hing wagers on which gender answers t he quest ion correct ly m ore oft en. Dat a regarding quest ion and regarding which gender answers correct ly m ore frequent ly are updat ed in real- t im e. Addit ional services were added such as elect ronic m ail, fax, not ificat ion service in which users are not ified when predet erm ined event s occur ( e.g., receipt of em ail, scheduled appoint m ent s) , int egrat ion of em ail int o t he SMS service, paging and it s int egrat ion int o t he SMS service, int eract ive banking, and inform at ion updat e services ( e.g., financial inform at ion, weat her, news, direct ory assist ance) . The SMS service on t he SMS- enabled phone is always on—t he phone is able t o receive or send a short m essage at any t im e, even if a voice conversat ion is in progress. This is possible since t he SMS m essages are delivered t o and from t he cell phones or wireless devices over t he syst em 's cont rol channel and not over t he voice channel. This helps in preserving t he voice qualit y while ut ilizing t he cont rol channel. However, t he cont rol channel capacit y im poses a lim it on t he lengt h of t he m essages since t he cont rol channel is used for cont rol services such as phone locat ion and call m anagem ent . As t he num bers of SMS m essages have increased, concerns about pot ent ial service problem s caused by overdem and have been m ent ioned. SMS guarant ees delivery of t he short m essage. Tem porary failures are ident ified, and t he short m essage is st ored in t he net work unt il t he dest inat ion device becom es available. The m essage is sent t o t he nearest Short Message Service Cent er ( SMSC) , which eit her delivers it t o t he addressed m obile device or forwards it t o t he next SMSC. The m essages are st ored and forwarded in t he SMSCs unt il t hey are received by t he addressees. I n t he delivery process, t he SMSC first sends an SMS request t o t he Hom e Locat ion Regist er ( HLR) t o locat e t he addressed m obile phone. The HLR is a dat abase used for st oring and m anaging subscript ions and service profiles. I f t he addressed m obile phone is found, t he HLR provides t he rout ing inform at ion. I f t he HLR cannot locat e t he addressed phone, t hen t he SMSC st ores t he m essage for a lim it ed t im e and t ries t o deliver t he m essage again when t he addressee connect s t o t he net work. I f t he addressee is found, t he m essage is delivered and t he SMSC receives verificat ion t hat t he m essage has been received by t he addressee. The sender is inform ed t hat t he SMS m essage has been received by t he addressee.

1.2.2 Enhanced Messaging Service (EMS) Enhanced Messaging Service ( EMS ) evolved from t he popular SMS. Service providers st art ed t o offer EMS in early 2000. EMS allows m uch richer cont ent opt ions t han t ext - based SMS m essages and can include pict ures, m elodies, sound m arks, graphic, anim at ions, font s, and form at t ed t ext . For exam ple, a person could send t his t ext t o his dat e wit h an SMS m essage " I love you." Using EMS, he could add a short anim at ion of flowers along wit h a love m elody. EMS uses t he st ore- and- forward m echanism provided by t he SMSCs. Sim ilar t o SMS, EMS uses t he cont rol channels. EMS is enabled by SMS concat enat ion, linking several short m essages t oget her. EMS support s basic and ext ended pict ures. Basic pict ures are black and whit e sm all ( 16x16 pixels) pict ures or large ( 32x32 pixels) pict ures. Ext ended pict ures can be, however, black and whit e, grayscale, or color. Ext ended pict ures can cont ain 255x255 pixels and can be t ransm it t ed in a com pressed form at . EMS also support s sound. There are a few predefined sounds, including low and high chim es and chords, Claps, TaDa, and drum sounds, as well as Not ify, User- Defined, and Ext ended Sounds. Predefined sounds are not t ransm it t ed over t he air. Only a reference t o t he sound is included in t he EMS. There are t en different sounds t hat can be added in t he m essage, and when t he sound reference is being displayed, t he referenced sound will be played. The sender can also download m elodies from various web sources. These sounds have t o be form at t ed according t o t he iMelody st andard. These m elodies can t ake up t o 128 byt es. Anim at ions in EMS provide users wit h a m uch st ronger alt ernat ive for expression t han plain pict ures. EMS support s predefined anim at ions t hat reflect happiness, sadness, flirt at iousness, gladness, skept icism , and grief. User- defined and ext ended anim at ions are support ed as well. Sim ilar t o sounds and pict ures, EMS predefined anim at ions are not sent as anim at ion over t he air. Only a reference t o t hem is included in t he EMS m essage. When t he m essage is received by t he addressee t he referenced anim at ion is displayed in a m anner t hat is specified by t he m anufact urer. User- defined anim at ions consist of four pict ures. There are t wo different anim at ion sizes: sm all ( 8x8 pixels) anim at ions and large ( 16x16 pixels) anim at ions. These anim at ions are sent over t he air int erface. Ext ended anim at ions m ay be black and whit e, grayscale, or color. The m axim um size of a single anim at ed fram e is 255x255 pixels. The repet it ion of t hese anim at ions m ay be cont rolled by t he sender and can be t ransm it t ed in a com pressed form . An EMS m essage can be sent t o a m obile phone or ot her wireless device, even if it does not support EMS, because all t he EMS com ponent s ( i.e., t ext form at t ing, pict ures, anim at ions, and sounds) are locat ed in t he m essage header. The EMS cont ent s included in t he header will be ignored by t he receiving m obile phone if it does not support EMS, and only t he t ext m essage will be displayed.

1.2.3 Multimedia Messaging Service (MMS ) Mult im edia Messaging Service ( MMS) t echnology is t he ult im at e m essaging applicat ion, allowing users t o creat e m essages t hat include any com binat ion of t ext , graphics, phot ographic im ages, speech, and audio or video clips. MMS support s st andard im age form at s such as JPEG and GI F, video form at s such as MPEG 4, and audio form at s such as MP3 and MI DI . Mult im edia m essaging depends on t he high t ransm ission speeds t hat will be available via 3G t echnologies.

MMS is expect ed t o be t he fut ure m ult im edia m essaging t echnology. As com pared wit h t he SMS size lim it at ion of 160 byt es, MMS will have no lim it at ion even if som e init ial im plem ent at ions m ay pose rest rict ions on t he m essage size. Rat her t han sending a sim ple m essage such as " I am lat e," t he user will be able t o send a m essage t hat bet t er explains t he reasons of being lat e. This unlim it ed m essage size will allow users t o express rich cont ent in t heir m essages for ult im at e expression of ideas, personalit y, and feelings. Being able t o send pict ures allows users t o share experiences wit h friends, fam ily m em bers, and business part ners. I n business applicat ions, users can capt ure relevant pict ures and send t hem t o colleagues or t o t heir hom e office or t hey can st ore t hem locally for fut ure ret rieval. MMS built - in present at ion layering can cont rol t im ing and synchronizat ion, allowing users t o view, list en t o, and read t he m essages sim ult aneously. MMS can support pict ure sharing. MMS will be able t o include st ill im ages such as pict ures, screensavers, post cards, graphics, greet ing cards, m aps, and business cards. I n addit ion, MMS will support anim at ion, video, cart oons, and int eract ive video. MMS will use t he Wireless Applicat ion Prot ocol ( WAP) as support ing t echnology and t he highspeed 2.5G and 3G t ransm ission t echnologies such as Enhanced Dat a rat e for Global Evolut ion ( EDGE) , General Packet Radio Service ( GPRS) , and Universal Mobile Telecom m unicat ions Syst em ( UMTS) . These high- speed connect ion t echnologies can provide users wit h t he necessary bandwidt h t o send and receive rich- cont ent m ult im edia m essages. I n SMS and EMS, t he m essages are t ransm it t ed over t he cont rol channel, which severely lim it s t he t ransm ission capacit y. I nst ead, MMS will use t he dat a channels used by all ot her voice and dat a applicat ions. Consequent ly, MMS will be able t o deliver m uch larger m essages. MMS is also expect ed t o use MExE ( Mobile Execut ion Environm ent ) , which is a flexible and secure applicat ion environm ent for 2.5G and 3G m obile devices. MExE includes a variet y of current t echnologies such as WAP and Java. Sim ilar t o SMS and EMS, MMS is a non- real- t im e service t hat rout es m ult im edia m essages t o MMS servers. MMS can include t he following: ●

●

●

●

Audio, sound, st ream ing audio, m elodies, songs. For exam ple, inst ead of sending a downloaded birt hday j ingle in EMS, a user can send a personalized video clip com bined wit h t he song " Happy Birt hday." I m ages or pict ures t hat are eit her st ored or t aken via a cam era t hat is at t ached or part of t he wireless phone or t erm inal. Users can t ake a snapshot and im m ediat ely send it t o an addressee. Video clips. I nit ially t he video clip m ay be lim it ed t o 30 seconds. This will allow users t o include video clips recorded wit h t heir digit al cam era and t o add t ext , labels, and appropriat e audio t o t heir m essages. MMS will also support st ream ing video so t hat users can subscribe t o news and ent ert ainm ent services and send t heir own capt ured video as a video st ream . Synchronized Mult im edia I nt egrat ion Language ( SMI L, pronounced " sm ile" ) which enables sim ple aut horing of int eract ive audiovisual present at ions. SMI L can be used t o creat e rich- cont ent m ult im edia present at ions t hat include st ream ing video and audio as well as pict ures, im ages, t ext , or any ot her m edium t ype. SMI L is based on an HTML- like language ( st andard defined by www.w3.org) .

1.2.4 Wireless Application Protocol (WAP) Wireless Applicat ion Prot ocol ( WAP) is an open, global st andard t hat provides a m icrobrowser environm ent opt im ized for wireless devices, such as phones and pagers. The goal is t o easily

access and int eract wit h inform at ion over t he I nt ernet . WAP is being em ployed in t he broad area of wireless applicat ions described in t his chapt er. This st andard is defined by t he WAP forum whose m em bers are wireless and I nt ernet com panies around t he world t o ensure int eroperabilit y and fost er growt h of t he wireless m arket s. I t ensures t hat all WAP- based applicat ions work across all devices, from cell phones t o m ore powerful hand- held devices. Also it allows for WAP- based applicat ions t o use less m em ory, processing power, display, and key handling t han legacy I nt ernet applicat ions. Thus t he WAP st andard helps in delivering t he m ult im edia cont ent while using less bandwidt h. The m ost recent st andard is WAP 2.0, published in 2002. I t adds support for t he st andard I nt ernet com m unicat ion prot ocols: I P, TCP, and HTTP. I t also allows applicat ions t o work over all exist ing and com ing wireless 3G t echnologies. I t provides a rich applicat ion environm ent , which enables delivery of inform at ion and int eract ive services t o digit al m obile phones, pagers, PDAs, and ot her wireless devices. I t addresses t he unique charact erist ics of wireless devices. These devices have hardware lim it at ions ( e.g., sm all screens, lim it ed bat t ery life, and lim it ed m em ory) requiring special at t ent ion t o user int erface design ( e.g., one- finger navigat ion) . WAP uses t he Pull Model, where t he user request s cont ent from t he server. WAP 2.0 adds t elephony support wit h WTA ( Wireless Telephony Applicat ion) which enables a wide range of advanced t elephony applicat ions in addit ion t o t heir legacy support of dat a- only funct ionalit y. These t ools include call handling services, such as m aking calls, answering t hem , placing t hem on hold, and redirect ing t hem . WAP 2.0 support s a Push Model, which allows server- based applicat ions t o send or " push" t he cont ent t o t he devices via a Push Proxy ( see Figure 1.9) . Push funct ionalit y is especially im port ant for sending news, announcem ent s, and not ificat ions t o int erest ed users. The cont ent can include st ock prices, locat ion- based prom ot ions, and t raffic updat e alert s. Wit hout push funct ionalit y, t hese t ypes of applicat ions would require t he devices t o specifically pull or request applicat ion servers for such new inform at ion. I n wireless environm ent s such pulling act ivit ies, if done persist ent ly and frequent ly, would burden t he lim it ed resource net work wit h wast eful t raffic.

Figu r e 1 .9 . Th e W AP Pr ogr a m m in g M ode l

The Ext ernal Funct ionalit y I nt erface ( EFI ) service specifies t he int erface bet ween t he Wireless Applicat ion Environm ent ( WAE) and plug- in m odules, which ext ends or enhances t he capabilit ies of browsers or ot her applicat ions. The EFI fram ework provides for fut ure growt h and ext endibilit y of support ed WAP devices and can be used t o access ext ernal devices ( e.g., sm art cards, GPS devices, digit al cam eras) . A WAP proxy ( or WAP gat eway) was required in t he original WAP st andard ( see Figure 1.10) . Such proxy handles t he conversion bet ween t he WAP client and t he origin server equipped wit h legacy I nt ernet servers t hat are not opt im ized for t he lim it ed bandwidt h of t he wireless channel. WAP 2.0 does not require a WAP proxy. However, use of a WAP proxy can provide several im port ant benefit s such as opt im izing t he com m unicat ion process in t he wireless net work and m ay offer m obile service enhancem ent s such as locat ion, secure channels ( privacy) , caching, and presence- based services. A proxy can also t ranslat e bet ween WAP and ot her WWW prot ocols, allowing WAP client s t o com m unicat e wit h servers t hat do not support WAP. A WAP proxy is necessary t o offer push funct ionalit y. Proxies m ay be locat ed in several locat ions, for exam ple, at t he wireless t elephone service providers and host ing com panies.

Figu r e 1 .1 0 . W AP Opt ion a l Pr ox y M ode l

< Day Day Up >

< Day Day Up >

Chapter 2. Quality of Service Fundamentals Sect ion 2.1. I nt roduct ion Sect ion 2.2. QoS Param et ers Sect ion 2.3. Mult im edia Applicat ion Requirem ent s Sect ion 2.4. QoS Services Sect ion 2.5. Realizat ion of QoS Services < Day Day Up >

< Day Day Up >

2.1 Introduction The previous chapt er int roduced a wide variet y of m ult im edia applicat ions used in wireless net works. These applicat ions em erge t o serve people's needs and som et im es creat e new services t hat at t ract people t o deploy wireless net works in t heir businesses or daily lives. Mult im edia applicat ions discussed in t his book m ainly focus on net work applicat ions, i.e., applicat ions where host s and m achines com m unicat e t hrough a net work. Unlike t he st and- alone m ult im edia applicat ions where t he m ult im edia cont ent s are originat ed and displayed on t he sam e m achine, t he net work m ult im edia applicat ion cont ent s t hat originat e on a source host are t ransm it t ed t hrough t he net work and displayed at t he dest inat ion host . Therefore, t here are a num ber of fact ors and com ponent s t hat affect t he perform ance of m ult im edia applicat ions such as: ●

●

●

●

Users: The ( hum an or nonhum an) ones who ut ilize t he m ult im edia applicat ions. Users' percept ion can influence t he evaluat ion of t he m ult im edia applicat ions' perform ance. Host m achine: The devices t hat operat e t he m ult im edia applicat ion ( source and dest inat ion host s) . The host s consist of a num ber of com ponent s such as processors, m edia st orage syst em s ( e.g., hard drive, CD- ROM) , display devices, and operat ing syst em s. Applicat ion: The st ruct ure or t he m echanism s built in t he m ult im edia applicat ion ( e.g., t he codec used in video com pression) . Net work: The net work com ponent s t hat t ransport t he m ult im edia cont ent s bet ween t he t wo host m achines ( source and dest inat ion) . Exam ples of hardware com ponent s include: swit ches, rout ers, net work int erface cards, gat eways, and firewalls. The net work also includes net work prot ocols t hat reside at each net work hardware elem ent .

Figure 2.1 shows a sim plified diagram of a com m unicat ion syst em t hat includes all t he com ponent s m ent ioned above. Host s connect t o t he net work t hrough t he net work int erface devices using various net work t echnologies ( e.g., I EEE 802.11 WLAN, Sat ellit e, Bluet oot h) . The net work cloud consist s of m ult iple segm ent s of int erconnect ed subnet works t hat est ablish t he com m unicat ion pat h bet ween t he host s. All of t hese com ponent s require m ult im edia support .

Figu r e 2 .1 . Com m u n ica t ion Syst e m

Mult im edia support issues can be present ed by using t he Qualit y of Service ( QoS) t erm , which is an overloaded t erm wit h various m eanings and perspect ives. There is lit t le consensus on t he precise definit ion of QoS. Different people and com m unit ies perceive and int erpret QoS in different ways. For exam ple, in t he net working com m unit y, QoS refers t o t he service qualit y or service level t hat t he net work offers t o applicat ions or users in t erm s of net work QoS param et ers, including lat ency or delay of packet s t raveling across t he net work, reliabilit y of packet t ransm ission, and t hroughput . However, in applicat ion com m unit ies, QoS generally refers t o t he applicat ion qualit y as perceived by t he user—t hat is, t he present at ion qualit y of t he video, t he responsiveness of int eract ive voice, and t he sound qualit y ( CD- like or FM- radiolike sound) of st ream ing audio. I nst ead of providing a precise QoS definit ion, we present a sim plified QoS m odel t hat includes t wo QoS perspect ives: Applicat ion/ User and Net works ( see Figure 2.2) .

Figu r e 2 .2 . QoS M ode l a n d Cor r e spon din g Se ct ion in Th is Ch a pt e r

We assum e t hat applicat ions and users are in t he sam e group because of t heir close relat ionship and t he com m on way t hey perceive qualit y. I n t his book we oft en use t he t erm s " applicat ions" and " users" int erchangeably. The applicat ions/ users expect a specific QoS in t erm s of response t im e, for exam ple consist ent perceived qualit y and unint errupt ed service. These QoS requirem ent s are passed t o t he net work im plicit ly or explicit ly and t he underlying net works are responsible in part for m eet ing t hese requirem ent s. Users are not concerned about how t he net work m anages it s resources or what m echanism s are involved in QoS provision. However, t he users are concerned about t he services t hat net works provide which direct ly im pact t he perceived qualit y of t he applicat ion. From t he net work perspect ive, t he net works' goal is t o provide t he QoS services t hat adequat ely m eet t he users' needs while m axim izing t he net work resources ( i.e., bandwidt h) ut ilizat ion. To achieve t his goal, t he net works analyze t he applicat ion requirem ent s, m anage t he net work resources, and deploy various net work QoS m echanism s. Alt hough t he QoS m odel here looks rat her sim ple, t here are a num ber of subt le issues t hat need t o be discussed. For exam ple: ● ●

● ● ● ●

What inform at ion is cont ained in t he applicat ion requirem ent s? Desired qualit y of t he applicat ion which relat es t o users' sat isfact ion is subj ect ive. How does t he desired qualit y m ap t o QoS param et ers which are m anaged by net works? What crit eria need t o be considered in t he m apping process? What are t he QoS param et ers? Which QoS m echanism s do t he net works use t o achieve diverse QoS support ? What kind of QoS service level can net works offer? What is t he relat ionship bet ween QoS services and applicat ion requirem ent s?

I n t he rem aining part of t his chapt er we at t em pt t o answer t hese issues. This chapt er is organized as follows. I n Sect ion 2.2 we int roduce t he applicat ion and net work QoS param et ers. Sect ion 2.3 discusses m ult im edia applicat ion requirem ent s in t erm s of QoS param et ers. QoS services are int roduced in Sect ion 2.4, and t he realizat ion of QoS services is described in

Sect ion 2.5. < Day Day Up >

< Day Day Up >

2.2 QoS Parameters The following QoS param et ers are relevant t o m ult im edia applicat ions: 1. Throughput or bandwidt h 2. Delay or lat ency 3. Delay variat ion ( delay j it t er) 4. Loss or error rat e

2.2.1 Throughput From t he applicat ion perspect ive, t hroughput refers t o t he dat a rat e ( bit s per second) generat ed by t he applicat ion. Throughput , m easured in t he num ber of bit s per second, som et im es is called bit rat e or bandwidt h. Bandwidt h is considered t o be t he net work resource t hat needs t o be properly m anaged and allocat ed t o applicat ions. The t hroughput required by an applicat ion depends on t he applicat ion charact erist ics. For exam ple, in a st ream ing video applicat ion, different video propert ies generat e different t hroughput . A user can select t he video qualit y by varying t he following video propert ies: ●

●

●

●

Fram e size: A funct ion of t he num ber of pixels in each row and colum n and of t he num ber of bit s per pixel. Fram e rat e: The refreshing video fram e rat e ( num ber of fram es per second) . Decreasing video fram e rat e reduces t he bandwidt h consum pt ion but com prom ises t he sm oot hness of t he video m ovem ent . Color dept h: The num ber of possible colors represent ed by a pixel. The 256- color video requires 8 bit s of dat a per pixel; whereas, t he 16- m illion- color video requires 24 bit s of dat a per pixel. Com pression: Reduct ion of t he bandwidt h consum pt ion at t he expense of im age qualit y. Exam ples of video com pression st andards include MPEG1, MPEG2, and MPEG4.

I n t he rem aining part of t his subsect ion we will describe t wo charact erist ics of an applicat ion dat a t raffic generat ion process: t he dat a t raffic generat ion rat e ( const ant and variable bit rat e) as well as dat a t raffic generat ion burst iness. We will show how t he t raffic generat ion process det erm ines t he applicat ions' required t hroughput .

2.2.1.1 Constant and Variable Bit Rate Const ant bit rat e ( CBR) applicat ions generat e dat a t raffic wit h const ant dat a rat e. Exam ples of CBR applicat ions include t he following: ●

●

Digit al t elephone Privat e Branch Exchange ( PBX) , which generat es 64 kbps const ant bit rat e Uncom pressed digit al video

Most of t he const ant bit rat e applicat ions are delay sensit ive and require const ant bandwidt h allocat ion. Allocat ing bandwidt h below t he required bandwidt h causes applicat ion failure. On t he ot her hand, allocat ing bandwidt h above t he requirem ent does not im prove t he user sat isfact ion, as shown in Figure 2.3.

Figu r e 2 .3 . Use r Sa t isfa ct ion a s a Fu n ct ion of Alloca t e d Ba n dw idt h for CBR a n d VBR t r a ffic

Variable bit rat e ( VBR) applicat ions generat e dat a t raffic wit h variable dat a rat e. The degree of bit rat e variabilit y depends on t he applicat ion. Exam ples of VBR applicat ions include t he following: ● ●

Com pressed video and audio Rem ot e login

VBR applicat ions require m inim um bandwidt h allocat ion in order t o operat e successfully. The m ore allocat ed bandwidt h, t he bet t er t he user- perceived qualit y. Bandwidt h allocat ion beyond t he m axim um required bandwidt h does not im prove user sat isfact ion as shown in Figure 2.3.

2.2.1.2 Burstiness The dat a t raffic burst iness m easures t he degree of bit rat e variabilit y of a VBR applicat ion. The burst iness is defined as t he rat io bet ween t he Mean Bit Rat e ( MBR) and Peak Bit Rat e ( PBR) where: ● ●

PBR is t he m axim um num ber of bit s in a short period of t im e MBR is t he average num ber of bit s in a long period of t im e

2.2.2 Delay Delay has a direct im pact on users' sat isfact ion. Real- t im e applicat ions require t he delivery of inform at ion from t he source t o t he dest inat ion wit hin a cert ain period of t im e. Long delays m ay cause incident s such as dat a m issing t he playback point , which in t urn reduces t he video fidelit y. Moreover, it can cause user frust rat ion during int eract ive t asks. When t he dat a t raffic is

carried across a series of com ponent s in t he com m unicat ion syst em t hat int erconnect s t he source and t he dest inat ion, each com ponent int roduces delay. We can cat egorize t he m ain sources of delay as follows: 1. Source- processing delay ( digit izat ion and packet izing delay) : This delay, which is int roduced by t he source t hat generat es t he packet s, depends on t he source host hardware configurat ion ( CPU power, RAM, m ot herboard, et c.) and it s current load ( e.g., t he num ber of applicat ions running sim ult aneously and t heir required hardware resources) . 2. Transm ission delay: The t ransm ission t im e of a packet is a funct ion of t he packet size and t ransm ission speed. 3. Net work delay: a. Propagat ion delay: The propagat ion delay from t he source t o t he dest inat ion is a funct ion of t he physical dist ance bet ween t he source and t he dest inat ion. b. Prot ocol delay: The delay is caused by t he com m unicat ion prot ocols execut ed at t he different net work com ponent s such as rout ers, gat eways, and net work int erface cards. The delay depends on t he prot ocols, t he load of t he net work, and t he configurat ion of t he hardware t hat execut es t he prot ocol. c. Out put queuing delay: The delay is caused by t he t im e a packet spends in t he out going link queue at a net work com ponent . For exam ple, such delay can be incurred at an int erm ediat e rout er out put queue. The delay depends on t he net work congest ion, t he configurat ion of t he hardware, and t he link speed. 4. Dest inat ion processing delay: This delay is int roduced by t he processing required at t he dest inat ion. For exam ple, such delay can be incurred in t he packet reconst ruct ion process. Sim ilar t o t he processing delay at t he source, t his delay depends on t he dest inat ion host hardware configurat ion and load. Figure 2.4 depict s t he end- t o- end delay diagram .

Figu r e 2 .4 . En d- t o- e n d D e la y D ia gr a m

2.2.3 Delay Variation Delay variat ion is a QoS m et ric t hat refers t o t he variat ion in t he delay int roduced by t he com ponent s along t he com m unicat ion pat h. Since each packet in t he net work t ravels t hrough different pat hs, and t he net work condit ions for each packet can be different , t he end- t o- end delay varies. For dat a generat ed at const ant rat e, t he delay j it t er dist ort s t he t im e synchronizat ion of t he original t raffic. As shown in Figure 2.5 t he packet s t ravel t hrough t he net work and experience different end- t o- end delays, reaching t he dest inat ion wit h t im ing dist ort ions ( incom plet e or delayed signal) relat ive t o t he original t raffic. There are several t echniques t o cope wit h delay j it t er at t he receiver end. For exam ple, in t echnique A ( Figure 2.5A) , t he receiver playbacks t he signal as soon as t he packet s arrive. The playback point is changed from t he original t im ing reference. This int roduces dist ort ion in t he playback signal. I n t echnique B ( Figure 2.5B) , t he receiver playbacks t he signal based on t he original t im ing reference. The lat e packet s t hat m iss t he playback point will be ignored. This also int roduces dist ort ion. I n t echnique C ( Figure 2.5C) a de- j it t ered buffer is used. All packet s will be st ored in t he buffer and held for som e t im e ( offset delay) before t hey are ret rieved by t he receiver wit h t he original t im ing reference. The fidelit y of t he signal will be m aint ained as long as t here are packet s available in t he buffer. Large delay j it t er requires large buffer space t o hold t he packet s and sm oot h out t he j it t er. A large buffer also int roduces large delays, which will be event ually const rained by t he applicat ion delay requirem ent . I n sum m ary, t here is a t radeoff bet ween t he following t hree fact ors: de- j it t ered buffer space, delay requirem ent , and fidelit y of t he playback signal.

Figu r e 2 .5 . D e la y Jit t e r a n d I t s Solu t ion s

2.2.4 Loss or Error Rate Packet loss direct ly affect s t he perceived qualit y of t he applicat ion. I t com prom ises t he int egrit y of t he dat a or disrupt s t he service. At t he net work level, packet loss can be caused by net work congest ion, which result s in dropped packet s. Anot her cause of loss is caused by bit errors t hat occur due t o a noisy com m unicat ion channel. Such loss will m ost likely occur in a wireless channel. There are several t echniques t o recover from packet loss or error such as packet ret ransm ission, error correct ion at t he physical layer, or codec at t he applicat ion layer t hat can com pensat e or conceal t he loss. < Day Day Up >

< Day Day Up >

2.3 Multimedia Application Requirements Different m ult im edia applicat ions have different QoS requirem ent s expressed in t erm s of t he following QoS param et ers as described in t he previous sect ion: t hroughput , delay, delay variat ion, and loss. I n m any cases, users can det erm ine t he applicat ion's QoS requirem ent s by invest igat ing t he fact ors t hat influence t he applicat ion qualit y ( i.e., t he t ask charact erist ics, user charact erist ics) . For exam ple, from experim ent at ion t hey conclude t hat for accept able qualit y, t he one- way delay requirem ent s of int eract ive voice should be less t han 250 m s. This delay value includes t he delay int roduced from all com ponent s in t he com m unicat ion pat h such as source delay, t ransm ission delay, net work delay, and dest inat ion delay. I n t his book, we focus on t he applicat ions' requirem ent s for m ult im edia support in wireless net works. Due t o advances in t he source and dest inat ion available processing power and t he very high- speed opt ical infrast ruct ure, we expect delay and bandwidt h bot t lenecks t o occur in t he wireless net work due t o t heir lim it ed available bandwidt h. Before present ing det ailed QoS requirem ent s for each applicat ion, we would like t o provide an overview of t he fact ors t hat influence t he applicat ion requirem ent s. Som e of t he fact ors are as follows: ● ●

●

Applicat ion int eract ivit y level: I nt eract ive and nonint eract ive applicat ions User/ Applicat ion charact erist ics: Delay t olerance and int olerance, adapt ive and nonadapt ive charact erist ics Applicat ion crit icalit y: Mission- crit ical and non- m ission- crit ical applicat ions

We will describe below each one of t hese fact ors.

2.3.1 Interactive and Noninteractive Applications An int eract ive applicat ion involves som e form of int eract ion ( act ion- react ion, request - response or exchange of inform at ion) bet ween t wo part ies ( people- t o- people, people- t o- m achine or m achine- t o- m achine) . Exam ples of int eract ive applicat ions include t he following: ● ●

●

People- t o- people applicat ion: I P t elephony, int eract ive voice/ video, videoconferencing People- t o- m achine applicat ion: Video- on- dem and ( VOD) , st ream ing audio/ video, virt ual realit y Machine- t o- m achine applicat ion: Aut om at ic m achine cont rol

The elapsed t im e bet ween int eract ions is essent ial t o t he success of an int eract ive applicat ion. The degree of int eract ivit y det erm ines t he level or st ringency of t he delay requirem ent . For exam ple, int eract ive voice applicat ions, which involve hum an int eract ion ( conversat ion) in real t im e, have st rict delay requirem ent s ( in t he order of m illiseconds) . St ream ing ( playback) video applicat ions involve less int eract ion, ( i.e., int eract ion m ost ly occurs during st art , st op, forward, or reverse act ion on t he video) and do not require real- t im e response. Therefore st ream ing applicat ions have m ore relaxed delay requirem ent s ( in t he order of seconds) . Oft en t he applicat ions' delay t olerance is det erm ined by t he users' delay t olerance ( i.e., higher delay t olerance leads t o m ore relaxed delay requirem ent s) . Delay j it t er is also relat ed t o QoS support for int eract ive t asks. As discussed above, t he delay

j it t er can be correct ed by de- j it t ering buffer t echniques. However, t he buffer int roduces delay in t he original signal, which event ually affect s t he int eract ivit y of t he t ask. I n general, an applicat ion wit h st rict delay requirem ent s also has st rict delay j it t er requirem ent s.

2.3.2 Tolerance and Intolerance Tolerance and int olerance describe t he users' sensit ivit y t o changes in QoS param et er values. We next describe users' t olerance t o lat ency and dist ort ion. ●

●

Lat ency t olerance and int olerance: This charact erist ic det erm ines t he st ringency of t he delay requirem ent . As we have discussed above, st ream ing m ult im edia applicat ions are m ore lat ency t olerant t han int eract ive m ult im edia applicat ions. The degree of lat ency t olerance depends on users sat isfact ion, users expect at ion, or t he urgency of t he applicat ion ( i.e., rem ot e m achine cont rol in t he m anufact uring product ion line is lat ency int olerant ) . Dist ort ion t olerance and int olerance: The t olerance t o t he fidelit y of t he applicat ion qualit y depends on fact ors such as users' sat isfact ion, users' expect at ion, and t he applicat ion m edia t ypes. For exam ple, users are m ore t olerant t o video dist ort ion t han t o audio dist ort ion. I n t his case, during congest ion, t he net work has t o m aint ain t he qualit y of t he audio out put over t he qualit y of t he video out put .

2.3.3 Adaptive and Nonadaptive Characteristics Adapt ive and nonadapt ive aspect s m ost ly describe t he m echanism s invoked by t he applicat ions t o adapt t o QoS degradat ion. The com m on adapt ive t echniques are rat e adapt at ion and delay adapt at ion: ●

●

Rat e adapt ive applicat ion can adj ust t he dat a rat e inj ect ed int o t he net work. During net work congest ion, t he applicat ions reduce t he dat a rat e by dropping som e packet s, increasing t he codec dat a com pression, or changing t he m ult im edia propert ies ( i.e., reducing t he resolut ion or color dept h of t he video) . These schem es will m ost likely cause degradat ion of t he perceived qualit y but will keep it wit hin accept able levels. Delay- t olerant adapt ive applicat ions t olerat e a cert ain level of delay j it t er by deploying t he de- j it t ered buffer or adapt ive playback t echnique ( Figure 2.5) .

Adapt at ion is t rigged by som e form of im plicit or explicit feedback from t he net work or end user. Again, like t olerance, adapt ive applicat ions provide som e room for net works t o readj ust t he QoS services.

2.3.4 Application Criticality Mission- crit ical and non- m ission- crit ical aspect s reflect t he im port ance of applicat ion usage, which det erm ines t he st rict ness of t he QoS requirem ent s. Failing t he m ission m ay result in disast rous consequences. For exam ple: ●

●

Rem ot e surgery: The surgeon perform s an operat ion t hrough rem ot e surgical equipm ent . Life and deat h of t he pat ient m ay depend on t he prom pt ness and accuracy of t he surgical equipm ent cont rol. Telem edicine: The accuracy of m edical im ages ( i.e., m agnet ic resonance im age, x- ray im age, ult rasound im age) is ext rem ely im port ant . Dist ort ed im ages m ay lead t o wrong diagnosis.

2.3.5 Representation of Application Requirements There are t wo com m on ways t o express t he applicat ions' requirem ent s: quant it at ive expression and qualit at ive expression.

2.3.5.1 Quantitative Expression Applicat ion requirem ent s are expressed in t erm s of QoS param et ers wit h quant ifiable values t hat can be det erm ined from t he applicat ion's t echnical specificat ions ( i.e., video codec—MPEG 1, MPEG2, MPEG4, HDTV) or from experim ent at ion. Som e applicat ions m ay obt ain t hese values using runt im e m easurem ent s. ●

Throughput : I t is m ost ly expressed as t he average dat a rat e: ❍ ❍ ❍ ❍

●

●

Uncom pressed HDTV: 1.5 Gbps MPEG4: 5 kbps – 4 Mbps I TU- T G.711: 64 kbps ETSI GSM 06.10: 16 kbps

For applicat ions t hat generat e VBR t raffic, t he average dat a rat e does not properly capt ure t he t raffic charact erist ics. Burst iness param et ers ( average rat e, peak rat e, m axim um burst size) m ay also be included. Delay and delay j it t er: The value is provided in t he form of a bound ( i.e., t he delay should be less t han a cert ain value) . I n m ission- crit ical or delay int olerant applicat ions, t he net work has t o follow t he delay requirem ent st rict ly. I n delay- t olerant applicat ions, t he bound value is t he average value. The net work can provide t he service in a m ore relaxed m anner ( i.e., som e of t he packet s can m iss t heir deadline) . Loss: I t is m ost ly expressed as a st at ist ical value ( i.e., t he percent age of lost packet s should be less t han a cert ain value) .

2.3.5.2 Qualitative Expression I n cont rast t o t he quant it at ive expression, QoS requirem ent s are expressed qualit at ively—for exam ple, " …. get t he bandwidt h as m uch as you can give… " or " … expect low lat ency…" or " …. provide lower lat ency t han cert ain applicat ions…." There are a num ber of reasons why som e applicat ions ( e.g., web browsing, rem ot e login) express t heir QoS requirem ent s qualit at ively: som e applicat ions only need bet t er service t han ot hers, or som e applicat ions are incapable of quant ifying t heir QoS requirem ent s, or it is t oo cost ly t o quant ify t he QoS requirem ent s. Since t hese applicat ions do not specify quant it at ive QoS requirem ent s, it is int ract able for net works t o provide quant it at ive services.

2.3.6 Examples of Application Requirements 2.3.6.1 Interactive Voice Requirem ent s for voice applicat ions such as voice over I P ( VoI P) can be exem plified as follows: ●

Bandwidt h: VoI P requires relat ively low bandwidt h ( i.e., 64 kbps in I TU G.711) and t he

t raffic pat t ern is relat ively const ant . Voice t raffic is packet ized in sm all packet s of around 44 t o 200 byt es result ing in short t ransm ission t im es. Since voice conversat ions cont ain up t o 60% silence, voice encoding algorit hm s yield lower bandwidt h requirem ent s wit h accept able qualit y. Table 2.1 shows t he speech coding st andard. Mean Opinion Score ( MOS) in Table 2.1 reflect s t he sound qualit y in a quant it at ive way. MOS is an em pirical value det erm ined by averaging t he opinion score ( i.e., t he range from 1= worst t o 5= best ) of a sam ple group of list eners who list en and evaluat e t he voice qualit y.

Ta ble 2 .1 . Spe e ch Codin g St a n da r ds

●

Code c

Ba n dw idt h ( k bps)

Sou n d Qu a lit y ( M OS)

Code c Com ple x it y

I TU- T G.711

64

> 4

—

I TU- T G.722

48- 64

3.8

low

I TU- T G.723.1

6.4/ 5.3

3.9

high

I TU- T G.726

32

3.8

low

I TU- T G.728

16

3.6

low

I TU- T G.729

8

3.9

m edium

GSM 06.10

13

3.5

low

GSM 06.20

5.6

3.5

high

GSM 06.60

12.2

> 4

high

GSM 06.70

4.8 – 12.2

> 4

high

Delay: Since VoI P involves hum an int eract ion ( conversat ion) , hum an percept ion is sensit ive t o bot h t he sound qualit y and t he conversat ion gap and response. Table 2.2 displays delay guidelines for VoI P.

Ta ble 2 .2 . D e la y Gu ide lin e s for VoI P On e - W a y D e la y

< 100 – 150 m s

●

Effe ct on Pe r ce ive d Qu a lit y

Excellent qualit y ( undet ect able delays)

150 – 250 m s

Accept able qualit y ( slight delays)

250 – 300 m s

Unaccept able qualit y

Delay j it t er: A large delay j it t er causes delayed packet s which im pact t he qualit y of t he voice ( conversat ion gap) . Table 2.3 shows t he delay j it t er guidelines for VoI P.

Ta ble 2 .3 . D e la y Jit t e r Gu ide lin e s for VoI P D e la y Jit t e r

Effe ct on Pe r ce ive d Qu a lit y

< 40 m s

●

Excellent qualit y—Undet ect able j it t er

40—75 m s

Accept able qualit y

Over 75 m s

Unaccept able qualit y

Not only t he packet loss has im pact on t he voice qualit y, but also t he pat t ern of t he packet loss has im pact . For exam ple, t he loss of one packet m ay not be not iceable due t o t he sophist icat ed codecs t hat can conceal t he loss. However, t wo or m ore consecut ive lost packet s can cause voice qualit y det eriorat ion.

2.3.6.2 Video Applications Video applicat ions can be classified int o t wo groups: int eract ive video ( i.e., video conferencing, long- dist ance learning, rem ot e surgery) and st ream ing video ( i.e., RealVideo, Microsoft ASF, QuickTim e, Video on Dem and, HDTV) . As shown in Table 2.4, video applicat ions' bandwidt h requirem ent s are relat ively high depending on t he video codec.

Ta ble 2 .4 . Vide o Code c Ba n dw idt h Re qu ir e m e n t s

●

●

Vide o Code c

Ba n dw idt h

Uncom pressed HDTV

1.5 Gbps

HDTV

360 Mbps

St andard definit ion TV ( SDTV)

270 Mbps

Com pressed MPEG2 4: 4: 4

25 – 60 Mbps

Broadcast qualit y HDTV ( MPEG 2)

19.4 Mbps

MPEG 2 SDTV

6 Mbps

MPEG 1

1.5 Mbps

MPEG 4

5 kbps – 4 Mbps

H.323 ( H.263)

28 kbps – 1 Mbps

I nt eract ive video applicat ions: I nt eract ive video which involves hum an int eract ion requires low end- t o- end delay. Low delay j it t er is also required especially in m issioncrit ical applicat ions such as rem ot e surgery. I nt eract ive video users are m ore t olerant t o video dist ort ion t han st ream ing video users. St ream ing video applicat ions: St ream ing video is considered t o be one- way com m unicat ion ( from video server t o users) . The only int eract ion bet ween t he server

and t he user occurs during st art , st op, forward, or reverse com m ands. Therefore, t he delay is not a m ain concern. However, t his applicat ion is delay j it t er int olerant . Large delay j it t er causes packet s t o m iss t he playback point , causing video dist ort ion. < Day Day Up >

< Day Day Up >

2.4 QoS Services Net works receive from t he applicat ions ( im plicit ly or explicit ly) t heir QoS requirem ent s t hrough quant it at ive or qualit at ive expression. Net works need t o respond t o t hese request s by supplying QoS services using a num ber of QoS m echanism s. I n t his sect ion, we describe various t ypes of QoS services and in t he next chapt er we will int roduce t he m echanism s required t o provide such services. I n t he rest of t he book we will analyze each wireless t echnology and discuss which kind of services it can provide and which kind of applicat ions it can support . We cat egorize t he QoS services as follows: 1. What kind of service is provided t o applicat ions: quant it at ive services, qualit at ive services, best effort services 2. To which ent it ies ( individual or group [ class] of applicat ions) t he net work provides service: per- flow QoS services, per- class QoS services. Net works m ay use a com binat ion of QoS services ( i.e., per- flow and quant it at ive, per- class and quant it at ive) . Som e net works m ay include m ult iple t ypes of QoS services in order t o support a wide range of applicat ions. Exam ple of QoS services include t he following: ●

● ●

Guarant eed I nt egrat ed Services ( I nt Serv) , which provide per- flow and quant it at ive QoS service Cont rolled Load I nt egrat ed Services, which provide per- flow and qualit at ive QoS service Different iat ed Services ( DiffServ) , which provide per- class and qualit at ive QoS service

We next describe in m ore det ail each one of t hese QoS services.

2.4.1 Quantitative (Guaranteed) Services Quant it at ive ( guarant eed) services, or services t hat deliver hard QoS, guarant ee t he provision of t he applicat ion quant it at ive requirem ent s. This service delivers t he highest qualit y of service. The guarant eed services guarant ee t he net work perform ance ( i.e., bandwidt h, delay, delay j it t er) in det erm inist ic or st at ist ical t erm s. For exam ple, net works guarant ee t he m inim um bandwidt h provided t o an applicat ion or guarant ee t he delay bound of packet delivery wit hin a cert ain value. The services are suit able for applicat ions t hat require quant it at ive perform ance guarant ee such as m ission- crit ical and int eract ive applicat ions. As described in t he next chapt er, a num ber of QoS m echanism s are required t o enable t hese services.

2.4.2 Qualitative (Differentiated) Services Qualit at ive ( different iat ed) services, or services t hat deliver soft QoS, provide relat ive services. I n ot her words, qualit at ive services m ay provide lower delay t o one class of applicat ions t han t o anot her class of applicat ions. One prom inent different iat ed services exam ple is t he priorit y service. An applicat ion t hat belongs t o a higher priorit y class will receive service before applicat ions t hat belong t o a lower priorit y class.

2.4.3 Best Effort Services Best effort services provide net work services wit hout any perform ance guarant ees. All t raffic is t reat ed equally. Best effort services are suit able for dat a t raffic ( i.e., FTP, em ail, web pages) t hat does not require m inim um bandwidt h or t im ed delivery.

2.4.4 Per-Flow QoS Services Per- flow QoS services provide service assurance t o individual flows ( applicat ions) quant it at ively or qualit at ively. For exam ple, int eract ive video has st rict er delay requirem ent s t han st ream ing video. Therefore, t he net work has t o provide different services t o each applicat ion in order t o m eet it s individual needs. Per- flow classificat ion ( QoS m echanism t hat different iat es t he flows) is essent ial t o t he im plem ent at ion of per- flow QoS services.

2.4.5 Per-Class QoS Services Applicat ions are cat egorized int o different groups ( classes) based on different crit eria ( i.e., QoS requirem ent , organizat ion, applicat ion t ypes, prot ocol fam ilies) . Per- class QoS services provide service assurance t o individual classes quant it at ively or qualit at ively. Applicat ions in t he sam e class will experience t he sam e QoS. The per- class classificat ion ( QoS m echanism t hat ident ifies and different iat es t he collect ive ent it ies) is essent ial t o t he im plem ent at ion of per- class QoS services. < Day Day Up >

< Day Day Up >

2.5 Realization of QoS Services There are a large num ber of approaches, m echanism s, and t echnologies deployed in t he net work in order t o enable t he QoS services int roduced in t he previous sect ion. The bandwidt h, which is t he m ain net work resource, needs t o be dist ribut ed t o all applicat ions in a way t hat sim ult aneously sat isfies all QoS requirem ent s. To enable QoS services, t here are t wo m ain approaches ( philosophies) based on how t hey deal wit h bandwidt h planning in order t o enable QoS services: ●

●

Bandwidt h over- provisioning: When t he current net work bandwidt h cannot provide QoS support , t he net work infrast ruct ure is upgraded. Using a higher bandwidt h infrast ruct ure m ay result in less congest ion and t herefore lower delivery delays. Bandwidt h m anagem ent : This approach proposes t o m anage t he bandwidt h using QoS m echanism s. Exam ples of QoS m echanism s which will be described in t he next chapt er are classificat ion, adm ission cont rol, resource reservat ion, channel access, packet scheduling, and policing.

There is an endless argum ent bet ween t he im plem ent at ion of t hese t wo approaches. I n wired net works, where bandwidt h is abundant , over- provisioning is t he winning approach. However, in wireless net works where t he bandwidt h is lim it ed ( spect rum is unavailable or is t oo expensive t o purchase) bandwidt h m anagem ent t echniques have t o be deployed. < Day Day Up >

< Day Day Up >

Chapter 3. QoS Mechanisms Sect ion 3.1. I nt roduct ion Sect ion 3.2. Classificat ion Sect ion 3.3. Channel Access Mechanism Sect ion 3.4. Packet Scheduling Mechanism s Sect ion 3.5. Traffic Policing Mechanism Sect ion 3.6. Resource Reservat ion Signaling Mechanism s Sect ion 3.7. Adm ission Cont rol Sect ion 3.8. QoS Archit ect ure < Day Day Up >

< Day Day Up >

3.1 Introduction I n t he previous chapt er, we int roduced t he fundam ent al QoS concept s. I n t his chapt er we int roduce a num ber of key QoS m echanism s t hat enable QoS services. At t he end of t his chapt er, we provide a general fram ework for analyzing t he QoS support of each wireless t echnology present ed in t he rest of t his book. QoS m echanism s can be cat egorized int o t wo groups based on how t he applicat ion t raffic is t reat ed: 1) t raffic handling m echanism s, and 2) bandwidt h m anagem ent m echanism s ( see Figure 3.1) .

Figu r e 3 .1 . QoS M e ch a n ism s in a W ir e le ss N e t w or k

Traffic handling m echanism s ( som et im es called I n- t raffic m echanism s) are m echanism s t hat classify, handle, police, and m onit or t he t raffic across t he net work. The m ain m echanism s are: 1) classificat ion, 2) channel access, 3) packet scheduling, and 4) t raffic policing. Bandwidt h m anagem ent m echanism s ( som et im es called Out - of- t raffic m echanism s) are m echanism s t hat m ange t he net work resources ( e.g., bandwidt h) by coordinat ing and configuring net work devices' ( i.e., host s, base st at ions, access point s) t raffic handling m echanism s. The m ain m echanism s are: 1) resource reservat ion signaling and 2) adm ission cont rol. < Day Day Up >

< Day Day Up >

3.2 Classification The lowest service level t hat a net work can provide is best effort service, which does not provide QoS support . I n best effort service, all t raffic is handled equally regardless of t he applicat ion or host t hat generat ed t he t raffic. However, som e applicat ions need QoS support , requiring bet t er t han best effort service such as different iat ed or guarant eed service. For a net work t o provide select ive services t o cert ain applicat ions, first of all, t he net work requires a classificat ion m echanism t hat can different iat e bet ween t he different applicat ions. The classificat ion m echanism ident ifies and separat es different t raffic int o flows or groups of flows ( aggregat ed flows or classes) . Therefore, each flow or each aggregat ed flow can be handled select ively. The classificat ion m echanism can be im plem ent ed in different net work devices ( i.e., end host s, int erm ediat e devices such as swit ches, rout ers, access point s) . Figure 3.2 shows a sim plified diagram of a classificat ion m odule t hat resides on an end host and on an int erm ediat ed device.

Figu r e 3 .2 . Cla ssifica t ion

Applicat ion t raffic ( at t he end host ) or incom ing t raffic from ot her host s ( at t he int erm ediat e device) is ident ified by t he classificat ion m echanism and is forwarded t o t he appropriat e queue await ing service from ot her m echanism s such as t he packet scheduler. The granularit y level of t he classificat ion m echanism can be per- user, per- flow, or per- class depending on t he t ype of QoS services provided. For exam ple, per- flow QoS service requires per- flow classificat ion while per- class QoS service requires per- class classificat ion. To ident ify and classify t he t raffic, t he t raffic classificat ion m echanism requires som e form of t agging or m arking of packet s. There are a num ber of t raffic classificat ion approaches. Som e of

approaches are suit able for end host s and som e for int erm ediat e host s. Figure 3.3 shows an exam ple of som e t raffic classificat ion approaches which are im plem ent ed in t he different Open Syst em I nt erconnect ion ( OSI ) layers.

Figu r e 3 .3 . Ex a m ple s of Ex ist in g Cla ssifica t ion on Ea ch OSI La ye r

3.2.1 Data Link Layer Classification Dat a link layer, or Layer 2, classifies t he t raffic based on t he t ag or field available in Layer 2 header. An exam ple of Layer 2 classificat ion is I EEE ( I nst it ut e of Elect rical and Elect ronics Engineers) 802 user priorit y. The I EEE 802 header includes a 3- bit priorit y field t hat enables eight priorit y classes. I t aim s t o support service different iat ion on a Layer 2 net work such as a LAN. The end host or int erm ediat e host associat es applicat ion t raffic wit h a class ( based on t he Policy, or t he service t hat t he applicat ion expect s t o receive) and t ags t he packet s' priorit y field in t he I EEE 802 header. A classificat ion m echanism ident ifies packet s by exam ining t he priorit y field of t he I EEE 802 header and forwards t he packet s t o t he appropriat ed queues. I EEE recom m ends m apping t he priorit y value and t he corresponding service as shown in Table 3.1.

Ta ble 3 .1 . Ex a m ple of M a ppin g be t w e e n Pr ior it y a n d Se r vice s Pr ior it y

Se r vice

0

Default , assum ed t o be best effort service

1

Less t han best effort service

2

Reserved

3

Reserved

4

Delay sensit ive, no bound

5

Delay sensit ive, 100m s bound

6

Delay sensit ive, 10m s bound

7

Net work cont rol

3.2.2 Network Layer Classification Net work layer, or Layer 3 classificat ion, classifies packet s using Layer 3 header. Layer 3 classificat ion enables service different iat ion in Layer 3 net work. An exam ple of Layer 3 classificat ion is I PTOS ( I nt ernet prot ocol t ype of service) , DSCP ( I nt ernet prot ocol different ial service code point ) . I Pv4 and I Pv6 st andard defined a priorit izat ion field in t he I P header which can be used for Layer 3 classificat ion. RFC 1349 defined a TOS field in I Pv4 header. The t ype of service field consist s of a 3- bit precedence subfield, a 4- bit TOS subfield, and t he final bit which is unused and is set t o be 0. The 4- bit TOS subfield enables 16 classes of service. I n I Pv6 header t here is an 8- bit class of service field ( see Figure 3.4) . Lat er t he I nt ernet Engineering Task Force ( I ETF) different iat ed services working group redefined I Pv4 I PTOS t o be DSCP, which is shown in Figure 3.4. DSCP has a 6- bit field enabling 64 classes of service.

Figu r e 3 .4 . St r u ct u r e of I PTOS a n d D SCP in I Pv4 a n d I Pv6

3.2.3 Transport Layer Classification (5-tuplet IP Header) A 5- t uplet I P header ( source I P, dest inat ion I P, source port , dest inat ion port , and prot ocol I P) can be used for t ransport layer classificat ion. A 5- t uplet I P header can uniquely ident ify t he individual applicat ion or flow. This classificat ion provides t he finest granularit y and support s perflow QoS service. However, t he 5- t uplet I P header classificat ion has som e lim it at ions: ●

●

I t is suit able for edge net works, but it is not suit able for core net works t hat carry very large am ount s of t raffic. Maint aining queues for each individual flow can be an overwhelm ing t ask. I f t he t raffic passes t hrough a firewall t hat uses NAT ( net work address t ranslat ion) , t he real I P address ( i.e., t he I P address of t he t raffic source) is hidden from net works out side t he firewall. Therefore, t he 5- t uplet I P header exposed t o a net work out side t he firewall cannot uniquely ident ify t he applicat ion.

3.2.4 Application or User Classification

The applicat ion or user can be uniquely ident ified by using user/ applicat ion ident ificat ion ( I D) . The I D assignm ent m ay be st at ic ( i.e., t he policy or t he cont ract ) or dynam ic ( i.e., connect ion signaling) . For t he connect ion signaling, t here is a cent ral st at ion or ent it y in t he net work t hat is responsible for m aking t he decision whet her t o allow a new session t o j oin t he net work. First , t he applicat ion or user sends t he connect ion request t o t he cent ral st at ion. Then, if t he new connect ion is adm it t ed, it will be assigned a unique I D num ber. Packet s from t he applicat ion will be associat ed wit h an I D num ber. < Day Day Up >

< Day Day Up >

3.3 Channel Access Mechanism I n wireless net works, all host s com m unicat e t hrough a shared wireless m edium . When m ult iple host s t ry t o t ransm it packet s on t he shared com m unicat ion channel, collisions can occur. Therefore, wireless net works need a channel access m echanism which cont rols t he access t o t he shared channel. There are t wo t ypes of channel access m echanism s: 1) collision- based channel access and 2) collision- free channel access. Each t ype of channel access m echanism can provide different QoS services.

3.3.1 Collision-Based Channel Access Collision- based channel access is a dist ribut ed channel access m et hod t hat provides m echanism s t o avoid collisions and t o resolve collisions in case t hey occur. A classic collisionbased channel access m echanism developed for wired LANs and im plem ent ed in Et hernet is CSMA/ CD ( Carrier Sense Mult iple Access wit h Collision Det ect ion) . I n collision- based channel access schem es, collisions can occur leading t o t he need for ret ransm issions. The collision probabilit y depends on t he num ber of act ive ( wit h packet s for t ransm ission) users in t he net work. High t raffic load increases t he num ber of collisions and ret ransm issions, increasing t he delay. Since we deal wit h st ochast ic t raffic, t he num ber of collisions and re- t ransm issions is random as well, leading t o an unbound delay. Therefore, collision- based channel access schem es can provide best effort service. All host s in t he net work receive equal bandwidt h and experience t he sam e unbounded delay. The service level can be im proved by: ●

●

Over- provisioning, whereby all t raffic will receive am ple of bandwidt h and experience low delay. Adding a priorit y schem e in t he collision- based channel access—t hat is, using different sized backoff windows for different priorit y classes. This will enable t he provision of different iat ed services. An exam ple of such a solut ion is described in t he proposed I EEE 802.11e ( Chapt er 4) .

Exist ing solut ions in wireless net works such as I EEE 802.11 DCF, Hom eRF use collision- based channel access prot ocols sim ilar t o Et hernet CSMA/ CD, denot ed as CSMA/ CA where CA st ands for Collision Avoidance.

3.3.2 Collision-Free Channel Access I n a collision- free channel access m echanism t he channel is arbit rat ed such t hat no collisions can occur. Only one host is allowed t o t ransm it packet s t o t he channel at any given t im e. Collision, t herefore, will not occur. Exam ples of collision- free channel access t echniques are polling and TDMA ( Tim e Division Mult iple Access) .

3.3.2.1 Polling A host in t he net work, or a specialized net work device such as an Access Point or Base St at ion, is designat ed as t he poller, which cont rols all access t o t he wireless channel by t he ot her host s denot ed as pollees. Pollees are not allowed t o t ransm it packet s unless t hey receive a polling packet from t he poller. Thus, t here is no collision. Som e pollees m ay receive t he poll m ore oft en t han ot hers. The polling frequency ( t he num ber of polls in a period of t im e) reflect s t he

bandwidt h allocat ion. A poller can dynam ically allocat e bandwidt h t o pollees by adj ust ing t he polling frequency dynam ically.

3.3.2.2 TDMA (Time Division Multiple Access) A TDMA schem e divides t he channel access opport unit y int o fram es and each fram e is divided int o t im e slot s. A host is allowed t o t ransm it packet s in a predefined t im e slot , as shown in Figure 3.5.

Figu r e 3 .5 . Tim e D ivision M u lt iple Acce ss ( TD M A) Sch e m e

The num ber of t im e slot s assigned t o a host per fram e reflect s t he bandwidt h allocat ed for t he host . This t echnique requires a m ast er host t hat is designat ed t o m anage t he t im e slot assignm ent for all t he host s in t he net work. This Mast er host det erm ines t he num ber of t im e slot s t hat each host will be allowed t o t ransm it and not ifies t he host s using som e signaling m echanism . There are a num ber of t im e slot assignm ent philosophies: ●

●

St at ic t im e slot assignm ent : Each host receives a fixed t im e slot assignm ent which can be provided during t he connect ion set up. Dynam ic t im e slot assignm ent : The t im e slot assignm ent changes dynam ically during t he lifet im e of t he session as a funct ion of t he t raffic load, applicat ion QoS requirem ent s, and channel condit ions. This slot assignm ent policy is m ore flexible and leads t o bet t er channel ut ilizat ion. However, it leads t o signaling overhead required t o com m unicat e t he slot assignm ent changes t o t he different host s.

Collision- free channel access schem es provide t ight channel access cont rol t hat can provide a t ight delay bound. Therefore, t hese schem es are good candidat es for QoS provision t o applicat ions wit h st rict QoS requirem ent s.

< Day Day Up >

< Day Day Up >

3.4 Packet Scheduling Mechanisms Packet scheduling is t he m echanism t hat select s a packet for t ransm ission from t he packet s wait ing in t he t ransm ission queue. I t decides which packet from which queue and st at ion are scheduled for t ransm ission in a cert ain period of t im e. Packet scheduling cont rols bandwidt h allocat ion t o st at ions, classes, and applicat ions. As shown in Figure 3.6, t here are t wo levels of packet scheduling m echanism s: 1. I nt rast at ion packet scheduling: The packet scheduling m echanism t hat ret rieves a packet from a queue wit hin t he sam e host . 2. I nt erst at ion packet scheduling: The packet scheduling m echanism t hat ret rieves a packet from a queue from different host s.

Figu r e 3 .6 . Pa ck e t Sch e du lin g

Packet scheduling can be im plem ent ed using hierarchical or flat approaches. ●

Hierarchical packet scheduling: Bandwidt h is allocat ed t o st at ions—t hat is, each st at ion

●

is allowed t o t ransm it at a cert ain period of t im e. The am ount of bandwidt h assigned t o each st at ion is cont rolled by int erst at ion policy and m odule. When a st at ion receives t he opport unit y t o t ransm it , t he int rast at ion packet scheduling m odule will decide which packet s t o t ransm it . This approach is scalable because int erst at ion packet scheduling m aint ains t he st at e by st at ion ( not by connect ion or applicat ion) . Overall bandwidt h is allocat ed based on st at ions ( in fact , t hey can be groups, depart m ent s, or com panies) . Then, st at ions will have t he aut horit y t o m anage or allocat e t heir own bandwidt h port ion t o applicat ions or classes wit hin t he host . Flat packet scheduling: Packet scheduling is based on all queues of all st at ions. Each queue receives individual service from t he net work.

Packet scheduling m echanism deals wit h how t o ret rieve packet s from queues, which is quit e sim ilar t o a queuing m echanism . Since in int rast at ion packet scheduling t he st at us of each queue in a st at ion is known, t he int rast at ion packet scheduling m echanism is virt ually ident ical t o a queuing m echanism . I nt erst at ion packet scheduling m echanism is slight ly different from a queuing m echanism because queues are dist ribut ed am ong host s and t here is no cent ral knowledge of t he st at us of each queue. Therefore, som e int erst at ion packet scheduling m echanism s require a signaling procedure t o coordinat e t he scheduling am ong host s. Because of t he sim ilarit ies bet ween packet scheduling and queuing m echanism s we int roduce a num ber of queuing schem es ( First I n First Out [ FI FO] , St rict Priorit y, and Weight Fair Queue [ WFQ] ) and briefly discuss how t hey support QoS services.

3.4.1 First In First Out (FIFO) First I n First Out ( FI FO) is t he sim plest queuing m echanism . All packet s are insert ed t o t he t ail of a single queue. Packet s are scheduled in order of t heir arrival. Figure 3.7 shows FI FO packet scheduling.

Figu r e 3 .7 . FI FO Pa ck e t Sch e du lin g

FI FO provides best effort service—t hat is, it does not provide service different iat ion in t erm s of bandwidt h and delay. The high bandwidt h flows will get a larger bandwidt h port ion t han t he low bandwidt h flows. I n general, all flows will experience t he sam e average delay. I f a flow increases it s bandwidt h aggressively, ot her flows will be affect ed by get t ing less bandwidt h, causing increased average packet delay for all flows. I t is possible t o im prove QoS support by adding 1) t raffic policing t o lim it t he rat e of each flow and 2) adm ission cont rol.

3.4.2 Strict Priority Queues are assigned a priorit y order. St rict priorit y packet scheduling schedules packet s based on t he assigned priorit y order. Packet s in higher priorit y queues always t ransm it before packet s in lower priorit y queues. A lower priorit y queue has a chance t o t ransm it packet s only when t here are no packet s wait ing in a higher priorit y queue. Figure 3.8 illust rat es t he st rict priorit y packet scheduling m echanism .

Figu r e 3 .8 . St r ict Pr ior it y Pa ck e t Sch e du lin g

St rict priorit y provides different iat ed services ( relat ive services) in bot h bandwidt h and delay. The highest priorit y queue always receives bandwidt h ( up t o t he t ot al bandwidt h) and t he lower priorit y queues receive t he rem aining bandwidt h. Therefore, higher priorit y queues always experience lower delay t han t he lower priorit y queues. Aggressive bandwidt h spending by t he high priorit y queues can st arve t he low priorit y queues. Again, it is possible t o im prove t he QoS support by adding 1) t raffic policing t o lim it t he rat e of each flow and 2) adm ission cont rol.

3.4.3 Weight Fair Queue (WFQ) Weight Fair Queue schedules packet s based on t he weight rat io of each queue. Weight , w i , is assigned t o each queue i according t o t he net work policy. For exam ple, t here are t hree queues A, B, C wit h weight s w 1 , w 2 , w 3 , respect ively. Queues A, B, and C receive t he following rat ios of available bandwidt h: w 1 / ( w 1 + w 2 + w 3 ) , w 2 / ( w 1 + w 2 + w 3 ) , and w 3 / ( w 1 + w 2 + w 3 ) , respect ively, as shown in Figure 3.9.

Figu r e 3 .9 . W e igh t Fa ir Qu e u e Pa ck e t Sch e du lin g

Bandwidt h abuse from a specific queue will not affect ot her queues. WFQ can provide t he required bandwidt h and t he delay perform ance is direct ly relat ed t o t he allocat ed bandwidt h. A queue wit h high bandwidt h allocat ion ( large weight ) will experience lower delay. This m ay lead t o som e m ism at ch bet ween t he bandwidt h and delay requirem ent s. Som e applicat ions m ay require low bandwidt h and low delay. I n t his case WFQ will allocat e high bandwidt h t o t hese applicat ions in order t o guarant ee t he low delay bound. Som e applicat ions m ay require high bandwidt h and high delay. WFQ st ill has t o allocat e high bandwidt h in order for t he applicat ions t o operat e. Of course, applicat ions will sat isfy t he delay but som et im es far beyond t heir needs. This m ism at ch can lead t o low bandwidt h ut ilizat ion. However, in real life, WFQ m ost ly schedules packet s t hat belong t o aggregat ed flows, groups, and classes ( inst ead of individual flows) where t he goal is t o provide link sharing am ong groups. I n t his case delay is of less concern. The elem ent ary queuing m echanism s int roduced above will be t he basis of a num ber of packet scheduling variat ions. Before we m ove our discussion t o t he next QoS m echanism s, it is wort h m ent ioning t hat in som e im plem ent at ions t he channel access m echanism and packet scheduling m echanism are not m ut ually exclusive. There is som e overlap bet ween t hese t wo m echanism s and som et im es t hey are blended int o one solut ion. When we discuss QoS support of each wireless t echnology in lat er chapt ers, in som e cases, we will discuss bot h m echanism s t oget her. < Day Day Up >

< Day Day Up >

3.5 Traffic Policing Mechanism Traffic policing is t he m echanism t hat m onit ors t he adm it t ed sessions' t raffic so t hat t he sessions do not violat e t heir QoS cont ract . The t raffic policing m echanism m akes sure t hat all t raffic t hat passes t hrough it will conform t o agreed t raffic param et ers. When violat ion is found ( e.g., m ore t raffic is sent t han was init ially agreed upon in t he QoS cont ract ) , a t raffic policing m echanism is enforced by shaping t he t raffic. Because t raffic policing shapes t he t raffic based on som e known quant it at ive t raffic param et ers, m ult im edia ( real- t im e) applicat ions are nat urally com pat ible t o t raffic policing. Most m ult im edia applicat ion t raffic ( voice, video) is generat ed by a st andard codec which generally provides cert ain knowledge of t he quant it at ive t raffic param et ers. Traffic policing can be applied t o individual m ult im edia flows. Non- real- t im e t raffic does not provide quant it at ive t raffic param et ers and usually dem ands bandwidt h as m uch as possible. Therefore, t raffic policing enforces non- real- t im e t raffic ( i.e., lim it s t he bandwidt h) based on t he net work policy. Such policing is usually enforced on aggregat ed nonreal- t im e flows. Traffic policing can be im plem ent ed on end host s or int erm ediat e host s. Exam ples of t raffic policing m echanism s include t he leaky bucket and t he t oken bucket .

3.5.1 Leaky Bucket The leaky bucket m echanism is usually used t o sm oot h t he burst iness of t he t raffic by lim it ing t he t raffic peak rat e and t he m axim um burst size. This m echanism , as it s nam e describes, uses t he analogy of a leaky bucket t o describe t he t raffic policing schem e. The bucket 's param et ers such as it s size and t he hole's size are analogous t o t he t raffic policing param et ers such as t he m axim um burst size and m axim um rat e, respect ively. The leaky bucket shapes t he t raffic wit h a m axim um rat e of up t o t he bucket rat e. The bucket size det erm ines t he m axim um burst size before t he leaky bucket st art s t o drop packet s. The m echanism works in t he following way. The arriving packet s are insert ed at t he t op of t he bucket . At t he bot t om of t he bucket , t here is a hole t hrough which t raffic can leak out at a m axim um rat e of r byt es per second. The bucket size is b byt es ( i.e., t he bucket can hold at m ost b byt es) . Let us follow t he leaky bucket operat ion by observing t he exam ple shown in Figure 3.10. We assum e first t hat t he bucket is em pt y. ●

●

●

Figure 3.10 ( A) : I ncom ing t raffic wit h rat e R which is less t han t he bucket rat e r. The out going t raffic rat e is equal t o R. I n t his case when we st art wit h an em pt y bucket , t he burst iness of t he incom ing t raffic is t he sam e as t he burst iness of t he out going t raffic as long as R < r. Figure 3.10 ( B) : I ncom ing t raffic wit h rat e R which is great er t han t he bucket rat e r. The out going t raffic rat e is equal t o r ( bucket rat e) . Figure 3.10 ( C) : Sam e as ( B) but t he bucket is full. Non- conform ant t raffic is eit her dropped or sent as best effort t raffic.

Figu r e 3 .1 0 . Le a k y Bu ck e t M e ch a n ism

3.5.2 Token Bucket The t oken bucket m echanism is alm ost t he sam e as t he leaky bucket m echanism but it preserves t he burst iness of t he t raffic. The t oken bucket of size b byt es is filled wit h t okens at rat e r ( byt es per second) . When a packet arrives, it ret rieves a t oken from t he t oken bucket ( given such a t oken is available) and t he packet is sent t o t he out going t raffic st ream . As long as t here are t okens in t he t oken bucket , t he out going t raffic rat e and pat t ern will be t he sam e as t he incom ing t raffic rat e and pat t ern. I f t he t oken bucket is em pt y, incom ing packet s have t o wait unt il t here are t okens available in t he bucket , and t hen t hey cont inue t o send. Figure 3.11 shows an exam ple of t he t oken bucket m echanism . ●

●

●

Figure 3.11 ( A) : The incom ing t raffic rat e is less t han t he t oken arrival rat e. I n t his case t he out going t raffic rat e is equal t o t he incom ing t raffic rat e. Figure 3.11 ( B) : The incom ing t raffic rat e is great er t han t he t oken arrival rat e. I n case t here are st ill t okens in t he bucket , t he out going t raffic rat e is equal t o t he incom ing t raffic rat e. Figure 3.11 ( C) : I f t he incom ing t raffic rat e is st ill great er t han t he t oken arrival rat e ( e. g., long t raffic burst ) , event ually all t he t okens will be exhaust ed. I n t his case t he incom ing t raffic has t o wait for t he new t okens t o arrive in order t o be able t o send out . Therefore, t he out going t raffic is lim it ed at t he t oken arrival rat e.

Figu r e 3 .1 1 . Tok e n Bu ck e t M e ch a n ism

The t oken bucket preserves t he burst iness of t he t raffic up t o t he m axim um burst size. The out going t raffic will m aint ain a m axim um average rat e equal t o t he t oken rat e, r. Therefore, t he t oken bucket is used t o cont rol t he average rat e of t he t raffic. I n pract ical t raffic policing, we use a com binat ion of t he t oken bucket and leaky bucket m echanism s connect ed in series ( t oken bucket , t hen leaky bucket ) . The t oken bucket enforces t he average dat a rat e t o be bound t o t oken bucket rat e while t he leaky bucket ( p) enforces t he peak dat a rat e t o be bound t o leaky bucket rat e. Traffic policing, in cooperat ion wit h ot her QoS m echanism s, enables QoS support . < Day Day Up >

< Day Day Up >

3.6 Resource Reservation Signaling Mechanisms The t raffic handling m echanism s ( classificat ion, channel access, packet scheduling, and t raffic policing) we already described enable QoS services in each device. However, coordinat ion bet ween devices is essent ial t o deliver end- t o- end QoS services. Resource reservat ion signaling m echanism s inform t he net work ent it ies on t he QoS requirem ent s of t he m ult im edia applicat ions using t he net work resources. The net work devices will use t his inform at ion t o m anage t he net work resources ( i.e., bandwidt h) in order t o accom m odat e such requirem ent s. The net work devices cont rol t he net work resources and provide QoS services by configuring t he t raffic handling m echanism s. Resource reservat ion can be applied t o individual flows or aggregat ed flows. Resource reservat ion closely cooperat es wit h t he adm ission cont rol m echanism s t hat will be described in a lat er sect ion. Figure 3.12 shows a schem at ic diagram t hat describes t he coordinat ion bet ween t hese m echanism s.

Figu r e 3 .1 2 . Re sou r ce Re se r va t ion M e ch a n ism

The resource reservat ion m echanism s include t he following funct ions: ●

●

●

Provision of resource reservat ion signaling t hat not ifies all devices along t he com m unicat ion pat h on t he m ult im edia applicat ions' QoS requirem ent s. Delivery of QoS requirem ent s t o t he adm ission cont rol m echanism t hat decides if t here are available resources t o m eet t he new request QoS requirem ent s. Not ificat ion of t he applicat ion regarding t he adm ission result .

Resource Reservat ion Prot ocol ( RSVP) is a well- known resource reservat ion signaling m echanism . RSVP operat es on t op of I P, in t he t ransport layer, so it is com pat ible wit h t he current TCP/ I P based m echanism s ( i.e., I Pv4, I P rout ing prot ocol, and I P m ult icast m echanism ) and can run across m ult iple net works. RSVP's m ain funct ionalit y is t o exchange QoS requirem ent inform at ion am ong t he source host , t he dest inat ion host , and int erm ediat e devices. Using t his inform at ion, each net work device will reserve t he proper resources and configure it s t raffic handling m echanism s in order t o provide t he required QoS service. Once t he reservat ion process is com plet e, t he sender host is allowed t o t ransm it dat a wit h an agreed t raffic profile. I f a device or net work elem ent on t he com m unicat ion pat h does not have enough resources t o accom m odat e t he t raffic, t he net work elem ent will not ify t he applicat ion t hat t he

net work cannot support t his QoS requirem ent . I n order t o achieve end- t o- end resource reservat ion, all t he net work elem ent s along t he pat h ( source host , dest inat ion host , and rout ers) need t o be RSVP- enabled. Originally, RSVP was designed for support ing per- flow reservat ion. Current ly it is ext ended t o support per- aggregat e reservat ion. < Day Day Up >

< Day Day Up >

3.7 Admission Control Adm ission cont rol is t he m echanism t hat m akes t he decision whet her t o allow a new session t o j oin t he net work. This m echanism will ensure t hat exist ing sessions' QoS will not be degraded and t he new session will be provided QoS support . I f t here are not enough net work resources t o accom m odat e t he new sessions, t he adm ission cont rol m echanism m ay eit her rej ect t he new session or adm it t he session while not ifying t he user t hat t he net work cannot provide t he required QoS. Adm ission cont rol and resource reservat ion signaling m echanism s closely cooperat e wit h each ot her. Bot h are im plem ent ed in t he sam e device. There are t wo adm ission cont rol approaches: ●

●

Explicit adm ission cont rol: This approach is based on explicit resource reservat ion. Applicat ions will send t he request t o j oin t he net work t hrough t he resource reservat ion signaling m echanism . The request t hat cont ains QoS param et ers is forwarded t o t he adm ission cont rol m echanism . The adm ission cont rol m echanism decides t o accept or rej ect t he applicat ion based on t he applicat ion's QoS requirem ent s, available resources, perform ance crit eria, and net work policy. I m plicit adm ission cont rol: There is no explicit resource reservat ion signaling. The adm ission cont rol m echanism relies on bandwidt h over- provisioning and t raffic cont rol ( i. e., t raffic policing) .

The locat ion of t he adm ission cont rol m echanism depends on t he net work archit ect ure. For exam ple, in case we have a wide area net work such as a high- speed backbone t hat consist s of a num ber of int erconnect ed rout ers, t he adm ission cont rol m echanism is im plem ent ed on each rout er. I n shared m edia net works, such as wireless net works, t here is a designat ed ent it y in t he net work ( e.g., st at ion, access point , gat eway, base st at ion) t hat host s t he adm ission cont rol agent . This agent is in charge of m aking adm ission cont rol decisions for t he ent ire wireless net work. This concept is sim ilar t o t he SBM ( subnet bandwidt h m anager) which serves as t he adm ission cont rol agent in 802 net works. I n ad hoc wireless net works, t he adm ission cont rol funct ionalit y can be dist ribut ed am ong all host s. I n infrast ruct ure wireless net works where all com m unicat ion passes t hrough t he access point or base st at ion, t he adm ission cont rol funct ionalit y can be im plem ent ed in t he access point or base st at ion. < Day Day Up >

< Day Day Up >

3.8 QoS Architecture This sect ion shows how all t he QoS m echanism s described in t he previous subsect ions are working t oget her t o provide QoS support . Different applicat ions t hat co- exist in t he sam e net work m ay require different com binat ions of QoS m echanism s such as: ●

●

●

Applicat ions wit h quant it at ive QoS requirem ent s: These applicat ions m ost ly require QoS guarant eed services. Therefore, explicit resource reservat ion and adm ission cont rol are needed. They also require st rict t raffic cont rol ( t raffic policing, packet scheduling, and channel access) . Applicat ions wit h qualit at ive QoS requirem ent s: These applicat ions require high QoS levels but do not provide quant it at ive QoS requirem ent s. I n t his case we can use resource reservat ion and adm ission cont rol. They also require t raffic handling which delivers different iat ed services. Best effort : There is no need for QoS guarant ees. The net work should reserve bandwidt h for such services. The am ount of reserved bandwidt h for best effort t raffic is det erm ined by t he net work policy.

The QoS archit ect ure which cont ains different QoS m echanism s is different for each net work t opology. We will focus on t he QoS archit ect ure for ad hoc and infrast ruct ure wireless net works.

3.8.1 QoS Architecture for Infrastructure Wireless Networks I n infrast ruct ure wireless net works, t here are t wo t ypes of st at ions: end st at ions ( host s) and a cent ral st at ion ( i.e., access point , base st at ion) . The cent ral st at ion regulat es all t he com m unicat ion in t he net work—t hat is, t here is no peer- t o- peer com m unicat ion t hat occurs direct ly bet ween t he host s. The t raffic from a source host is sent t o t he cent ral st at ion and t hen t he cent ral st at ion forwards t he t raffic t o t he dest inat ion host . All t raffic handling ( classificat ion, t raffic policing, packet scheduling, and channel access) and resource reservat ion m echanism s reside in all st at ions ( end host s and cent ral st at ion) . I n addit ion, t he cent ral st at ion also includes an adm ission cont rol m echanism . Figure 3.13 shows a QoS archit ect ure for an infrast ruct ure wireless net work.

Figu r e 3 .1 3 . QoS Ar ch it e ct u r e of a n I n fr a st r u ct u r e W ir e le ss N e t w or k

There are som e variat ions in t he signaling m echanism s t hat configure t he t raffic handling m echanism s in each st at ion. We will point out t hese differences in each wireless t echnology chapt er.

3.8.2 QoS Architecture For Ad Hoc Wireless Networks All host s est ablish peer- t o- peer com m unicat ion in t he shared wireless m edia environm ent . All t raffic handling and resource reservat ion m echanism s reside in all host s. One of t he host s ( eit her a dedicat ed or a regular end host ) will be designat ed t o serve as an adm ission cont rol agent ( i.e., designat ed SBM [ DSBM] ) . Figure 3.14 shows a QoS archit ect ure for an ad hoc wireless net work.

Figu r e 3 .1 4 . QoS Ar ch it e ct u r e for a n Ad H oc W ir e le ss N e t w or k

< Day Day Up >

< Day Day Up >

Part 2: Wireless Local Area Networks

< Day Day Up >

< Day Day Up >

Chapter 4. IEEE 802.11 Sect ion 4.1. I EEE 802.11 Sect ion 4.2. I EEE 802.11e ( QoS Ext ension) < Day Day Up >

< Day Day Up >

4.1 IEEE 802.11 As shown in Figure 4.1, t he I EEE ( I nst it ut e of Elect rical and Elect ronics Engineers) st andards com m it t ee has defined a fam ily of I EEE 802 LAN ( local area net work) and MAN ( m et ropolit an area net work) st andards in which I EEE 802.11 defines t he wireless LAN ( WLAN) st andard wit h m ult iple physical layer opt ions including speeds of up t o 54 Mbps. The organizat ion is expect ed t o int roduce enhancem ent s t o t he st andard t hat will support QoS and ot her radio opt ions. Many com panies ( e.g., Cisco, 3Com , Linksys, D- Link) have int roduced com m unicat ion cards based on t his st andard.

Figu r e 4 .1 . I EEE 8 0 2 St a n da r ds

Com pat ibilit y and int eroperabilit y am ong various I EEE 802.11 vendors are m anaged by t he Wireless Et hernet Com pat ibilit y Alliance ( WECA) . Their cert ificat ion is referred as Wi- Fi, which st ands for Wireless Fidelit y. Wi- Fi cert ificat ion gives consum ers and businesses t he assurance t hat WLAN product s bearing t he Wi- Fi logo are com pat ible and work t oget her even when m anufact ured by different vendors. Such PC product s include PCMCI A ( PC Mem ory Card I nt ernat ional Associat ion) cards for not ebooks, PCI ( peripheral com ponent int erconnect / int erface) cards for deskt ops, and cards wit h USB ( Universal Serial Bus) int erface t hat can be used wit h eit her one. The wireless m edia are fundam ent ally different from t he wired m edia in t he following aspect s: ●

●

●

●

The wireless m edia have no clear physical boundaries in which users can t ransm it and receive dat a. Users can som et im es be covered by m ult iple WLANs. This fact requires a resolut ion m echanism t hat will enable t he users t o conduct effect ive com m unicat ion wit h m inim al int errupt ions from ot her WLANs' part icipant s. Users are suscept ive t o radio int erferences com ing from sources such as m icrowaves, causing t he wireless m edia t o be significant ly less reliable t hen t he wired m edia. Users are m obile—t hat is, t hey can change t heir locat ions, and roam wit hin t heir coverage areas. This m obilit y causes a dynam ic net work t opology, which leads t o t he need t o have procedures for t he users t o associat e/ disassociat e wit h t he specific WLAN.

●

●

Moreover, a user can be t em porarily hidden from som e ot her net work m em bers. The power lim it at ion, t he wireless m edia nat ure, and t he t opographic charact erist ics of t he com m unicat ion area im pose a lim it ed range for t he effect ive geographical coverage. This fact m ay require t he inst allat ion of several overlapping wireless LANs t o cover a larger geographical area. Many, if not m ost , of t he wireless devices are bat t ery powered. Hence, power saving is essent ial t o increase t he t im e bet ween bat t ery charges. This fact m ay cause sit uat ions in which t he user is t em porarily wit hout service.

The I EEE 802.11 fam ily consist s of a few st andards. The first st andard t hat was approved in 1997 det ails t he Wireless LAN Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions t hat support dat a rat es of 1 Mbps and 2 Mbps over t he 2.4 GHz I ndust rial, Scient ific, and Medical ( I SM) frequency band using eit her Frequency Hopping Spread Spect rum ( FHSS) or Direct Sequence Spread Spect rum ( DSSS) as t heir radio t echnologies as well as infrared. ( Not e: I SM frequency bands are nonlicensed frequency bands used for wireless net work, cordless phone, and ot her devices. I SM frequency bands include 902–928 MHz, 2.4– 2.5 GHz and 5.725–5.875 GHz.) The st andard det ails operat ional specificat ions for wireless connect ivit y for fixed, ad hoc, and m obile st at ions wit hin a local geographical area. This first st andard was int roduced aft er years of careful developm ent and has becom e popular wit h vendors who im plem ent ed significant part s of t his st andard. The I EEE 802.11 st andard cont ains t wo addit ions int roduced lat er: I EEE 802.11a enhances t he speed t o up t o 54 Mbps using Ort hogonal Frequency Division Mult iplexing ( OFDM) radio t echnology in t he 5 GHz band, and I EEE 802.11b enhances t he speed t o 5.5 Mbps and 11 Mbps using Com plem ent ary Code Keying ( CCK) m odulat ion in t he 2.4 GHz band. These st andards allow for m ult irat e support and backwards com pat ibilit y ( i.e., I EEE 802.11b support s speeds of 1, 2, 5.5, and 11 Mbps) . As long as users in t he sam e WLAN use t he sam e basic radio t echnology, even t hough wit h different dat a rat es, t hese users will be able t o com m unicat e using t he lowest speed. For exam ple, a lower speed DSSS based radio ( i.e., I EEE 802.11) will be able t o com m unicat e wit h a m ore advanced CCK based radio ( i.e., I EEE 802.11b) sharing t he lower speed ( 1 or 2 Mbps) of t he t wo radios. The I EEE 802.11g com m it t ee is working on t he definit ion of a st andard in which higher speeds of m ore t han 20 Mbps will be used wit hin t he 2.4 GHz band. This pot ent ial ext ension will be backwards com pat ible and int eroperable wit h I EEE 802.11b. I EEE 802.11g provides speeds of 1, 2, 5.5, 6, 9, 11, 12, 18, 24, 36, 48, and 54 Mbps. The I EEE 802.11 st andards consider dat a t raffic but lack proper considerat ion for m ult im edia applicat ions' QoS needs. Current ly, t he I EEE 802.11 group is working on an ext ension, I EEE 802.11e, t hat will specify how signaling and support for QoS can be achieved. Furt herm ore, t he I EEE 802.11 group is also working on a num ber of st andard supplem ent s as sum m arized in Table 4.1.

Ta ble 4 .1 . I EEE 8 0 2 .1 1 St a n da r d Su pple m e n t s St a n da r d Su pple m e n t s

Br ie f D e scr ipt ion

I EEE 802.11a

Define a PHY in newly allocat ed UNI I band. The st andard has been com plet ed and published as 8802- 11: 1999 ( E) / Am d 1: 2000 ( I SO/ I EC) ( I EEE St d. 802.11a- 1999 Edit ion) .

I EEE 802.11b

Define a high rat e PHY in t he 2.4 GHz band. The st andard has been com plet ed and published as I EEE St d. 802.11b- 1999.

I EEE 802.11c

Define MAC procedure for t he bridge operat ion. The st andard has been com plet ed and is now part of I EEE 802.11c st andard.

I EEE 802.11d

Define PHY requirem ent s t o ext end t he operat ion of 802.11 WLAN t o new regulat ory dom ains ( count ries) . The st andard has been com plet ed and published as I EEE St d. 802.11d- 2001.

I EEE 802.11e

Enhance current 802.11 MAC t o expand support for applicat ions wit h QoS requirem ent .

I EEE 802.11f

Recom m end pract ice for I nt er Access Point Prot ocol ( I APP) .

I EEE 802.11g

St andard for high- speed ( 20+ Mbps) PHY ext ensions t o t he 802.11b st andard.

I EEE 802.11h

Enhance 802.11 MAC and 802.11a PHY wit h net work m anagem ent , cont rol ext ensions for spect rum and t ransm it power m anagem ent in 5GHz license exem pt bands.

I EEE 802.11i

Enhance t he 802.11 MAC t o enhance securit y and aut hent icat ion m echanism s.

I EEE 802.11j

Enhance t he current 802.11 MAC and 802.11a PHY t o addit ionally operat e in newly available Japanese 4.9 GHz and 5 GHz bands.

I EEE 802.11k

Define Radio Resource Measurem ent enhancem ent s t o provide int erfaces t o higher layers for radio and net work m easurem ent s.

I EEE 802.11a, b, c, and d have already been approved, whereas I EEE 802.11e, f, g, h, i, j , and k are under developm ent .

4.1.1 Architecture 4.1.1.1 Network Topologies The st andard addresses t wo t opologies: ad hoc t opology, referred t o as I ndependent Basic Service Set ( I BSS) , and infrast ruct ure t opology, referred t o as Basic Service Set ( BSS) . A t opology t hat com bines several BSS cells is referred t o as Ext ended Service Set ( ESS) .

4.1.1.1.1 Ad Hoc Network As shown in Figure 4.2, an ad hoc net work or an I BSS consist s of st at ions wit hin m ut ual com m unicat ion range of each ot her via t he wireless m edium . Such a net work is creat ed spont aneously, wit hout preplanning, for ad hoc t em porary sit uat ions wit h lim it ed needs t o access t he I nt ernet . The I BSS is t he m ost basic t ype of an I EEE 802.11 WLAN and m ay cont ain only t wo st at ions. Figure 4.2 shows t wo I BSSs, each wit h t wo st at ions. I f a st at ion m oves out of it s I BSS, m eaning out of range, it can no longer com m unicat e wit h t he ot her I BSS m em bers.

Figu r e 4 .2 . Ad H oc Topology: I n de pe n de n t Ba sic Se r vice Se t s

4.1.1.1.2 Infrastructure Network The infrast ruct ure net work or BSS includes an access point ( AP) in addit ion t o t he st at ions. This AP act s as t he BSS arbit rat or, m eaning t hat t he AP will handle all t he BSS t raffic. The BSS t raffic can be eit her int ernal t raffic ( i.e., am ong t he BSS part icipant s) or ext ernal t raffic ( bet ween t he BSS part icipant s and out side t he BSS) . The AP int egrat es t he BSS wit hin t he dist ribut ion net work. For exam ple, all t raffic bet ween t he BSS part icipant s and t he I nt ernet will be delivered t hrough t he AP. Figure 4.3 shows an ESS com posed of t wo BSSs, each wit h t wo st at ions and an AP. Each BSS is int erconnect ed t o t he dist ribut ion syst em ( DS) ( which m ay connect t o t he I nt ernet ) t hrough t he AP. The key concept is t hat t he ESS net work appears as a single net work ent it y t o t he upper layers including t he applicat ions.

Figu r e 4 .3 . I n fr a st r u ct u r e Topology: Ex t e n de d Se r vice Se t

To allow effect ive com m unicat ion for users who const ant ly m ove from one BSS geographical area t o anot her BSS area, t he BSSs should be physically overlapping. I n t he st andard t here is no lim it t o t he dist ance bet ween t he BSSs and it is left up t o t he user or net work inst aller t o det erm ine t he BSS locat ion. Moreover, ot her users m ay decide t o inst all an I BSS in t he sam e geographical area. I n t his case, m ult iple WLANs will be collocat ed and funct ion effect ively if t he respect ive users coordinat e t he used frequencies.

4.1.1.1.3 Collocation Wireless net works can be locat ed in overlapping geographical areas ( see Figure 4.4) . Radio signals m ay propagat e from one net work t o anot her. A change in t he t ransm it t er posit ion and geographical layout m ay have a profound effect on t he propagat ion charact erist ics. Hence such net works do not have a clear boundary. The possibilit y of having m ult iple collocat ed operat ional WLANs not only allows m ult iple independent user groups t o work sim ult aneously in t he sam e place but also aggregat es and increases t he WLAN capacit y.

Figu r e 4 .4 . Colloca t e d W LAN s

Collocat ion is possible if net work inst allers m ake an effort t o coordinat e t he frequencies used by t heir respect ive WLANs. I f t hey use DSSS based WLANs t hey can coordinat e t he cent er frequency. Several cent ers are possible according t o t he region in which t he net works are inst alled. For exam ple, I EEE 802.11b in Nort h Am erica det ails t hree non- overlapping channels and six overlapping channels ( see Figure 4.5) . This m eans t hat up t o t hree net works can be locat ed wit hout m ut ual int errupt ions.

Figu r e 4 .5 . N or t h Am e r ica n Ch a n n e ls

Collocat ion is also possible in FHSS- based WLANs. I EEE 802.11b defines 78 hopping sequences ( each wit h 79 hops) grouped in t hree set s of 26 sequences each. This m eans t hat t heoret ically 26 FHSS WLAN syst em s can be collocat ed. However, pract ically fewer user groups can be collocat ed because of int erference from ot her groups. This num ber is est im at ed t o be 15. I n I EEE 802.11a collocat ion is possible if different WLANs use different operat ing frequencies. I n t he U.S., t here are t hree 100 MHz unlicensed nat ional inform at ion infrast ruct ure ( U- NI I ) bands: Lower U- NI I ( 5.15 t o 5.25 GHz) , Middle U- NI I ( 5.25 t o 5.35 GHz) , and Upper U- NI I ( 5.725 t o 5.825 GHz) .

4.1.1.2 Protocol Stack As shown in Figure 4.6, t he I EEE 802.11- 1997 base st andard focuses on t he following t wo part s

of t he prot ocol st ack: t he MAC part of t he dat a link layer and t he physical layer.

Figu r e 4 .6 . Por t ion s of t h e OSI Pr ot ocol St a ck Cove r e d by I EEE 8 0 2 .1 1 Ba se St a n da r d ( PLCP [ ph ysica l la ye r con ve r ge n ce pr ot ocol] ) . Abbr e via t ion s: M LM E—M AC Su bla ye r M a n a ge m e n t En t it y; PLM E— Ph ysica l La ye r M a n a ge m e n t En t it y; SAP—Se r vice Acce ss Poin t .

4.1.2 Physical Layer I EEE 802.11 allows for various wireless t echnologies: infrared and radio DSSS, FHSS, and OFDM. One t echnology cannot work wit h t he ot her. I n ot her words, radio FHSS cannot com m unicat e wit h a radio t hat em ploys DSSS. The st andard, t hough, specifies backwards com pat ibilit y for radios t hat use t he sam e t ransm ission t echnology. I n ot her words, a higher speed radio will be able t o com m unicat e wit h a lower speed radio via adj ust ing t he speed of t he higher speed radio t o t hat of t he lower speed radio. 802.11a allows for higher speeds in t he 5 GHz band based on OFDM: 6, 9, 12, 18, 24, 36, 48, and 54 Mbps ( support of 6, 12, and 24 Mbps dat a rat es is m andat ory) . OFDM is also considered by I EEE 802.11g at t he 2.4 GHz frequency range t o achieve speeds of 1, 2, 5.5, 6, 9, 11, 12, 18, 24, 36, 48, and 54 Mbps. The various radio flavors use slight ly different frequencies in different count ries and cont inent s. This m eans t hat a U.S. radio will not work properly wit h a Japanese radio. Ongoing work is conduct ed t o allow bet t er int eroperabilit y and t o m inim ize t he need for equipm ent m anufact urers t o produce a wide variet y of count ry- specific product s and for users t o t ravel wit h a bag full of count ry- specific WLAN cards. Such effort is reflect ed in t he I EEE 802.11d t hat supplem ent s t he I EEE 802.11 wit h feat ures t hat allow WLANs t o operat e wit hin t he rules of different count ries. Anot her effort is I EEE 802.11f, whose goal is t o achieve radio access point int eroperabilit y wit hin a m ult ivendor WLAN environm ent .

4.1.3 Media Access Control (MAC)

The MAC governs t he st at ions' access t o t he shared wireless m edium . The MAC which is locat ed wit h t he dat a link layer has a crucial role in providing QoS support t o users, especially when execut ing m ult im edia applicat ions. The MAC archit ect ure det ails t wo operat ional m odes t hat coexist ( see Figure 4.7) : t he Dist ribut ed Coordinat ion Funct ion ( DCF) and t he Point Coordinat ion Funct ion ( PCF) . The t wo m odes use a cycle st ruct ure ( see Figure 4.8) denot ed collision- free period repet it ion int erval, in which t he first t im e period ( called Cont ent ion- free Period or CFP) is governed by t he PCF m ode and t he second t im e period ( called Cont ent ion Period or CP) is governed by t he DCF m ode. DCF is a sim plist ic m ode t hat allows cont ent ion and collision of t raffic bet ween st at ions. Therefore, DCF is suit able for applicat ions t hat do not require QoS. PCF is a m ore com plex prot ocol t hat uses t he Point Coordinat or ( PC) which resides at t he AP. The PC arbit rat es t he channel access using a polling based approach. PCF is geared t o provide QoS support for m ult im edia applicat ions. However, t here are no known product s t hat im plem ent t he PCF m ode. QoS provisioning is going t hrough an ext ensive review by t he I EEE 802.11 organizat ion under t he ext ension I EEE 802.11e. This is discussed in Sect ion 2 of t his chapt er.

Figu r e 4 .7 . I EEE 8 0 2 .1 1 M AC Ar ch it e ct u r e

Figu r e 4 .8 . I EEE 8 0 2 .1 1 Collision - Fr e e Pe r iod Re pe t it ion I n t e r va l

I n addit ion t o defining t he m edia access rules for each m ode, t he st andard also defines procedures for t raffic fragm ent at ion and defragm ent at ion, m ult irat e support , aut hent icat ion and privacy. Fragm ent at ion, or slicing t he original dat a fram e int o sm aller t ransm it t ed fram es, is carried in order t o reduce t he loss of t raffic which can occur due t o collisions wit h ot her users' t raffic or due t o radio int erference. The sm aller t he fram e, t he less wireless m edia bandwidt h is lost in t he collision. However, sm aller fram es lead t o m ore overhead consist ing of fram es headers t hat cont ain addresses and ot her pert inent inform at ion. Hence, t here is a t radeoff which will det erm ine how m uch fragm ent at ion should be carried out . The st andard does not quant ify t he level of fragm ent at ion and leaves it t o t he im plem ent ers' discret ion.

4.1.3.1 Distributed Coordination Function (DCF) The fundam ent al channel access m echanism of t he I EEE 802.11 MAC is DCF, also known as Carrier Sense Mult iple Access wit h Collision Avoidance ( CSMA/ CA) . This is a random access m echanism t hat enables sharing of t he wireless m edia bet ween com pat ible radio t ransm it t ers. The st andard m andat es t hat t he DCF is im plem ent ed in all st at ions for use wit hin bot h t he I BSS and BSS configurat ions. I n t he DCF m ode a st at ion cannot t ransm it arbit rarily, but rat her needs t o go t hrough a set of st eps t o det erm ine whet her it can t ransm it . The goal of t hese st eps is t o reduce t he chance of collision wit h ot her st at ions' packet s. Such collision causes t he dest ruct ion of t he t ransm it t ed packet along wit h ot her st at ions' packet s. Before t ransm ission, t he user's t ransm it t er senses t he wireless m edium t o det erm ine if anot her st at ion is t ransm it t ing. I f t he wireless m edium is det erm ined t o be free, t he t ransm it t er will sense t he channel and wait before t ransm ission for a cert ain t im e referred t o as I nt erfram e Space ( I FS) . Once t he dat a are successfully received by t he addressee, t he addressee is required t o respond wit h an im m ediat e posit ive acknowledgm ent ( ACK) such t hat t he user is assured t hat t he dat a have arrived. I n case t he ACK is not received, t he user st at ion will go t hrough a process of ret ransm ission unt il an ACK is received. When a ret ransm ission is required, t he t ransm it t er needs t o wait an addit ional random backoff t im e before t ransm it t ing in order t o m inim ize collisions wit h t he ot her st at ions t hat also part icipat ed in t he collision. I EEE 802.11 defines four I FSs wit h different durat ion ( see Figure 4.9) : Short I nt erfram e Space ( SI FS) , PCF I nt erfram e Space ( PI FS) , DCF I nt erfram e Space ( DI FS) , and Ext ended I nt erfram e Space ( EI FS) . The durat ion of t hese int ervals is det erm ined in t he st andard as a funct ion of t he radio t echnology used. The durat ion of t hese int ervals det erm ines t he priorit y of t he packet s ( i. e., in case t he packet wait s SI FS, which is short est int erfram e int erval, t he packet has t he highest priorit y) .

Figu r e 4 .9 . I EEE 8 0 2 .1 1 I n t e r fr a m e Spa cin g ( I FS)

The I FS int ervals are used as follows: ●

● ●

SI FS is used as an int erfram e space bet ween a dat a fram e and it s ACK fram e, bet ween an RTS ( request t o send) fram e and it s CTS ( clear t o send) fram e bet ween fragm ent s of t he original dat a fram e, and by a st at ion responding t o a cont rol m essage in t he PCF m ode. PI FS is used in t he PCF m ode t o gain priorit y access t o t he m edium . DI FS is used in t he DCF m ode t o t ransm it dat a and m anagem ent fram es. I f t he wireless m edium is free for t he DI FS period t he st at ion is allowed t o t ransm it .

Despit e t his elaborat e m echanism collisions can occur if st at ions are far away from each ot her and cannot list en t o each ot her's t ransm issions or if t he st at ions st art t ransm it t ing concurrent ly. The m ore st at ions t here are in t he BSS or I BSS t he m ore collisions occur, and hence bandwidt h ut ilizat ion is det eriorat ed. The st andard provides a m echanism t o furt her m inim ize collisions by reserving t he channel for a cert ain period of t im e. During t his period of t im e only one st at ion is allowed t o t ransm it while ot hers defer. This reservat ion is obt ained t hrough t he following process. The t ransm it t ing st at ion ( i.e., t he st at ion t hat would like t o reserve t he channel) t ransm it s a special announcem ent t o all st at ions, announcing it s int ent t o t ransm it in t he period m ent ioned in t his announcem ent . The addressed st at ion is expect ed t o respond and acknowledge t his announcem ent . The announcem ent is referred t o as Request To Send ( RTS) packet and t he addressee's response is referred t o as Clear To Send ( CTS) packet . All st at ions who receive eit her t he RTS or CTS will refrain from t heir t ransm ission at t em pt s unt il t he end of t he t im e period announced in eit her t he RTS or CTS packet s. The RTS/ CTS m echanism cannot be used for broadcast and m ult icast t raffic. The RTS/ CTS burdens t he net work wit h addit ional t raffic and overhead t hat reduces som e of it s benefit . The st andard does not define when and in what sit uat ions t he RTS and CTS exchange should be carried out . This decision is left t o t he im plem ent er and even t he user in cases where t he wireless com m unicat ion card provides access t o t his decision. Based on t he CSMA/ CA prot ocol, each st at ion m aint ains a predict ion of t im e in which fut ure t raffic will occur. This predict ion m echanism is referred in t he st andard as t he Net work Allocat ion Vect or ( NAV) . This durat ion is announced eit her in t he RTS/ CTS packet s or in t he header of t he packet s sent during t he DCF m ode. The st at ions use t he NAV t o defer from accessing t he wireless m edium .

4.1.3.2 Point Coordination Function (PCF)

The I EEE 802.11 MAC st andard defined an opt ional access m et hod called Point Coordinat ion Funct ion ( PCF) which operat es during t he Cont ent ion- Free Period ( CFP) . This m ode can be used only in an infrast ruct ure net work and uses t he Point Coordinat or ( PC) as t he net work coordinat or or arbit er. I n a pract ical im plem ent at ion, t he PC can be included in t he AP. Using a cent ralized polling based approach, t he PC det erm ines which st at ion has t he right t o t ransm it and polls t his st at ion. A st at ion can t ransm it only when polled by t he PC. When polled, t he st at ion can t ransm it only one dat a fram e which can be eit her a new dat a packet or a ret ransm ission packet of a previous packet for which no ACK has been received. The st at ion is allowed, t hough, t o piggyback an ACK on t he t ransm it t ed dat a fram e. Thus, t he PCF provides a cont ent ion- free dat a t ransfer. To gain access t o t he m edium and st art t he PCF period ( in CFP) , t he PCF m ode uses t he PI FS int erval ( which is short er in durat ion t han t he DI FS int erval) along wit h special m anagem ent fram es t hat include t he durat ion of t he net work allocat ion vect or ( NAV) t o be set by all ot her st at ions. Using PI FS, t he PC has t he highest channel access priorit y. The PC sends t he NAV value in t he Beacon m essage. All st at ions set t heir NAV values which prevent t hem from accessing t he channel during t he CFP, t herefore prevent ing collisions. Thus, PCF provides collision- free access t o t he channel which has t he pot ent ial t o provide QoS support . As discussed lat er in t his chapt er, t he st andard does not provide t he polling int elligence in t erm s of when t o poll each st at ion. I t is left for t he im plem ent er t o decide t he polling int elligence.

4.1.4 Physical Layer Convergence Protocol (PLCP) The PLCP layer locat ed beneat h t he MAC has several responsibilit ies. I t carries out t he Carrier Sensing ( CS) of t he wireless m edia t o det erm ine whet her t here are ongoing t ransm issions on t he channel. I t also synchronizes t he speed of t he com m unicat ing st at ions' t ransm it t er and receiver and est ablishes t he com m unicat ion. To accom plish t hese responsibilit ies it adds addit ional inform at ion t o each packet . The CS m echanism can be configured wit h several Clear Channel Assessm ent ( CCA) opt ions: ●

●

●

CCA Mode 1: Energy above t hreshold. CCA report s a busy m edium upon det ect ing any energy above t he t hreshold. CCA Mode 4: Carrier sense wit h t im er. CCA st art s a t im er whose durat ion is defined in t he st andard and report s a busy m edium only upon t he det ect ion of a signal unt il t he t im er expires. Ot herwise it report s t he m edia is not busy. CCA Mode 5: A com binat ion of carrier sense and energy above t hreshold.

The synchronizat ion is obt ained by using packet headers which are t ransm it t ed in increasing speeds. This m eans t hat t he first part of t he header is t ransm it t ed in a low speed, t he second in a higher speed, and so fort h. This m echanism allows speed synchronizat ion but also slows down higher speed radios by as m uch as 30% .

4.1.5 QoS Support As we m ent ioned previously, I EEE 802.11 MAC has t wo m odes of operat ions, PCF and DCF, which deliver different QoS support . I n order t o clearly underst and t he QoS support , we first exam ine t he QoS m echanism s provided by I EEE 802.11.

4.1.5.1 QoS in IEEE 802.11 DCF QoS m echanism s ( Figure 4.10) 1. Classificat ion: There is no classificat ion in DCF. Therefore, no service different iat ion is provided. 2. Channel access: DCF uses cont ent ion- based m edia access cont rol prot ocol ( CSMA/ CA) . 3. Packet scheduling: The int rast at ion packet scheduling uses FI FO m echanism ( i.e., all t raffic in a st at ion is queued and t ransm it t ed in a first - in- first - out order) .

Figu r e 4 .1 0 . I EEE 8 0 2 .1 1 D CF QoS M e ch a n ism s

QoS service and support ed applicat ions DCF delivers best effort QoS service level which is suit able for non- real t im e applicat ions. There is no service different iat ion and no service guarant ee in t erm s of bandwidt h and delay.

4.1.5.2 QoS in IEEE 802.11 PCF QoS Mechanism s ( Figure 4.11) 1. Classificat ion: There is no classificat ion in PCF. Therefore, t here is no service different iat ion wit hin st at ions ( i.e., all t raffic in t he sam e st at ion is t reat ed equally) .

2. Channel access: PCF uses a polling based m edia access cont rol prot ocol. A st at ion is allowed t o t ransm it packet s only when it receives a polling m essage from t he PC. Since no collisions occur, PCF can deliver a predict able service perform ance. 3. Packet scheduling: Due t o t he lack of a classificat ion m echanism wit hin each st at ion, t he int rast at ion packet scheduling uses a FI FO m echanism . The int erst at ion packet scheduling is direct ly relat ed t o t he polling sequence and polling frequency ( t he num ber of polls t o a st at ion in a t im e period) which are not specified by t he st andard.

Figu r e 4 .1 1 . I EEE 8 0 2 .1 1 PCF QoS M e ch a n ism s

QoS service and support ed applicat ions PCF can deliver a cert ain level of guarant eed QoS service which is suit able for real- t im e applicat ions. To achieve guarant eed QoS service, t he following funct ions have t o be explicit ly det erm ined: ●

●

●

Algorit hm s t o det erm ine t he polling sequence and polling frequency t o properly allocat e bandwidt h and sat isfy applicat ions' delay requirem ent s. Algorit hm s t o det erm ine t he durat ion of t he CFP repet it ion int erval, CFP int erval, and CP int erval so t hat applicat ions' delay requirem ent s are sat isfied. QoS signaling: t o det erm ine t he proper polling sequence and polling frequency, t he PC needs t o obt ain from t he st at ions inform at ion on t he t raffic dynam ics at each st at ion as well as applicat ions' quant it at ive QoS requirem ent s. Cert ain QoS signaling is included in I EEE 802.11e.

< Day Day Up >

< Day Day Up >

4.2 IEEE 802.11e (QoS Extension) Realizing t he short com ings of t he QoS support wit hin t he current I EEE 802.11 MAC, t he I EEE organizat ion ( I EEE 802.11 Task Group E) has been involved in an ext ensive effort t o specify MAC enhancem ent s for bet t er QoS support . This effort and pot ent ial ext ension t o t he st andard are referred t o as I EEE 802.11e. This ext ension has not been approved yet . Hence t his discussion provides j ust a sim plified explanat ion and glim pse int o t he proposals and ideas discussed wit hin t his effort . The proposed I EEE 802.11e applies t o t he physical layers defined in I EEE 802.11 and I EEE 802.11a, b, and g. I n addit ion, t he proposal allows for QoS support even when legacy I EEE 802.11 st at ions are present . This is an im port ant feat ure t hat allows for friendly coexist ence of new I EEE 802.11e com pliant st at ions wit h legacy I EEE 802.11 st at ions. Several problem s wit h t he PCF m ot ivat ed t his effort . Such problem s include t he unpredict able delays associat ed wit h beacons due t o t he CSMA/ CA and PI FS m echanism s, unknown t ransm ission durat ions of t he polled st at ions caused by unknown fram e lengt hs and m odulat ion t echniques, and t he inabilit y t o different iat e am ong t raffic st ream s which can enable t he provision of higher priorit ies t o QoS sensit ive applicat ions. These problem s soft en t he QoS support in t he following ways: 1) t im e delay is unbounded causing delay sensit ive t ransm issions t o be queued for a lengt hy t im e, and 2) t here is an inabilit y t o provide t he needed bandwidt h guarant ees for bandwidt h sensit ive applicat ions. I EEE 802.11e's goal is t o deliver t he following: ●

●

Quant it at ive QoS services ( param et erized QoS) : I t provides QoS services t hat m eet t he applicat ion's quant it at ive requirem ent s in t erm s of t raffic specificat ions ( TSPEC) . Different iat ed services ( priorit ized QoS) : I t provides priorit y services am ong t raffic st ream s.

I EEE 802.11e proposes t he following QoS enhancem ent s: ● ●

● ●

Traffic classificat ion: t raffic cat egory and t raffic st ream Channel access and packet scheduling: enhanced MAC funct ions such as Enhanced DCF ( EDCF) and Hybrid Coordinat ion Funct ion ( HCF) QoS signaling New fram es for QoS support

4.2.1 Traffic Classification 802.11e provides t wo t ypes of classificat ion: t raffic cat egory ( TC) and t raffic st ream ( TS) .

4.2.1.1 Traffic Category A t raffic cat egory provides a t ool for applicat ions t o set a dist inct priorit y relat ive t o ot her dat a t o be t ransm it t ed over t he wireless m edia ( priorit ized QoS) . 802.11e defines t raffic cat egories t o support different iat ed services t o at m ost eight delivery priorit ies designat ed 0 t hrough 7. Dat a st ream s are classified t o one of t he eight t raffic cat egories. By default , user priorit y TC7 is m apped t o t he highest delivery priorit y and user priorit y TC1 is m apped t o t he lowest delivery priorit y ( see Table 4.2) . I t is possible t o have m ult iple dat a st ream s m apped t o t he sam e t raffic cat egory. Therefore, classificat ion based on t raffic cat egory is com parable t o per- class

classificat ion defined in Chapt er 3. Traffic Cat egory cooperat es wit h EDCF.

Ta ble 4 .2 . I EEE 8 0 2 .1 1 e Tr a ffic Ca t e gor ie s Pr ior it y Le ve l Use r Pr ior it y Tr a ffic Ca t e gor ie s Acr on ym

Lowest

Tr a ffic Type

1

TC1

BK

Background

2

TC2

--

Spare

0 ( default )

TC0

BE

Best effort

3

TC3

EE

Excellent effort

4

TC4

CL

Cont rolled load

5

TC5

VI

Video < 100 m sec delay and j it t er

6

TC6

VO

Voice < 10m sec delay and j it t er

7

TC7

NC

Net work cont rol

| | | | | | | | | | | | | Highest To m anage t hese t raffic cat egories, I EEE 802.11e has t o im plem ent a num ber of physical queues. Table 4.3 shows a m apping bet ween t he t raffic cat egories and t he queues.

Ta ble 4 .3 . I EEE 8 0 2 .1 1 e M a ppin g of Tr a ffic Ca t e gor y t o Qu e u e s

N u zm be r of Qu e u e s

D e fin in g Tr a ffic Type

1

BE

2

BE

3

BE

VO

4

BK

BE

5

BK

BE

6

BK

BE

7

BK

8

BK

--

CL

VO

CL

VO

CL

VI

VO

EE

CL

VI

VO

BE

EE

CL

VI

VO

NC

BE

EE

CL

VI

VO

NC

4.2.1.2 Traffic Stream To support quant it at ive QoS services ( param et erized QoS) , I EEE 802.11e defines t raffic st ream s ( TSs) . I n each st at ion t here are at m ost eight t raffic st ream s. Each t raffic st ream is associat ed wit h it s t raffic specificat ion ( TSPEC) which includes: 1. Quant it at ive obj ect ives for t raffic at t ribut es such as packet size and arrival rat es 2. Traffic charact erist ics ( const ant vs. variable dat a rat e, m axim um delivery delay, m axim um delay variance, et c.) 3. Acknowledgm ent policy Each dat a flow wit h quant it at ive QoS requirem ent is direct ly classified t o a t raffic st ream . Classificat ion based on t raffic st ream s is com parable t o per- flow classificat ion defined in Chapt er 3. Traffic cat egory cooperat es wit h HCF. Figure 4.12 shows t he classificat ion ( bot h t raffic cat egory and t raffic st ream ) in a st at ion.

Figu r e 4 .1 2 . I EEE 8 0 2 .1 1 e Cla ssifica t ion M e ch a n ism s

4.2.2 Channel Access and Packet Scheduling 802.11e provides enhanced MAC funct ions ( EDCF, HCF) which include bot h channel access and packet scheduling m echanism s. Therefore, we discuss bot h QoS m echanism s at once for each MAC. I EEE 802.11e defines t he following MAC prot ocol wit h t wo m odes of operat ion: t he Hybrid Coordinat ion Funct ion ( HCF) t o support QoS net work configurat ions and Enhanced DCF ( EDCF) . Sim ilar t o 802.11, 802.11e proposal has t wo phases of MAC operat ion: Cont ent ion Period ( CP) and Cont ent ion- Free Period ( CFP) , which alt ernat e over t im e cont inuously. I n each phase t he st at ion can send dat a packet s or t heir fragm ent s, according t o rules set by each phase. The EDCF is used in t he CP phase only, while t he HCF is used in bot h phases ( see Figure 4.13) .

Figu r e 4 .1 3 . I EEE8 0 2 .1 1 e CFP Re pe t it ion I n t e r va l

The HCF com bines funct ions from t he legacy DCF and PCF wit h som e enhanced, QoS- specific funct ions and fram e subt ypes. Such enhancem ent s allow a uniform set of fram e exchange sequences t o be used for QoS t ransfers during bot h t he CP and CFP. The HCF uses a cont ent ion based m edia access m et hod, referred as enhanced DCF ( EDCF) . St at ions m ay obt ain t ransm ission opport unit ies using one or bot h of t hese m edia access m et hods. St at ions t hat support t he 802.11e are referred as QoS St at ions ( QSTAs) . The QoS st at ion has an opt ion t o funct ion as t he cent ralized cont roller for all ot her st at ions wit hin t he WLAN or QoS support ing Basic Service Set ( QBSS) . This cent ralized cont roller is called Hybrid Coordinat or ( HC) or Point Coordinat or ( sim ilar t o 802.11 PCF) . The HC is expect ed t o be inst alled in t he 802.11e Access Point ( AP) .

4.2.2.1 Enhanced Distributed Coordination Function (EDCF) The EDCF provides different iat ed, dist ribut ed access t o t he wireless m edium for at m ost eight delivery priorit ies as defined above. EDCF channel access is derived from CSMA/ CA used in DCF wit h t he addit ion of priorit ies. Mult iple packet s are delivered t hrough m ult iple backoff inst ances wit hin one st at ion. Each backoff inst ance is labeled wit h a TC specific param et er. I n t he CP phase, each TC wit hin t he st at ions cont ends for t ransm ission and independent ly st art s a backoff process congruent t o t he CSMA/ CA aft er det ect ing t he channel being idle for an Arbit rat ion I nt erfram e Space ( AI FS) . The AI FS is at least DI FS, and can be increased for each TC. Consequent ly, dat a st ream s wit h higher AI FS values, m eaning longer backoff t im es, have lower priorit y access t o t he wireless m edia. Dat a st ream s wit h lower AI FS value will be able t o access t he wireless m edia before dat a st ream s wit h higher AI FS values ( see Figure 4.14) .

Figu r e 4 .1 4 . I EEE 8 0 2 .1 1 ED CF

Aft er t he QoS st at ions wait for AI FS, each backoff set s a count er t o a random num ber t hat is less t han t he size of t he cont ent ion window ( CW) . Each t raffic cat egory has it s own CW param et er. To achieve com pat ibilit y over legacy 802.11 WLANs, AI FS is set equal t o DI FS. Sim ilar t o DCF, when t he wireless m edium is det erm ined busy before t he count er reaches zero, t he backoff has t o wait for t he wireless m edium t o be idle for anot her AI FS period. I f an unsuccessful t ransm ission occurred, a new enlarged CW is com put ed by using t he persist ence fact or ( PF) . For exam ple, if PF= 1, CW rem ains unchanged. I f PF= 2, CW applies a binary exponent ial backoff algorit hm . Therefore, t he key param et ers ( m aint ained in each t raffic cat egory) t hat enable priorit ies are AI FS, CW, and PF. The t ransm it t ing st at ions can ret ain cont rol of t he wireless m edium by using t he short int erfram e spaces ( SI FSs) and t he virt ual carrier sense m echanism ( NAV) . The st at ion can t ransm it a num ber of packet s t hat will look like a single inst ance of act ivit y t o ot her cont ending st at ions. A QoS st at ion has t he opt ion t o im plem ent up t o eight t ransm ission queues based on TCs t hat dict at e t he queue priorit y. I n case t here are t wo or m ore TCs in t he st at ion and som e of t heir backoff count ers reach zero at t he sam e t im e, t he st at ion's scheduler grant s t he t ransm ission opport unit y t o t he queue wit h t he highest TC priorit y. EDCF can st ill result in collisions am ong st at ions t hat carry t raffic wit h t he sam e priorit y. A st at ion is allowed a m axim um t ransm ission durat ion ( t ransm ission opport unit y [ TXOP] ) as defined in t he

Beacon m essage. The RTS and CTS m echanism defined in 802.11 DCF can also be im plem ent ed in t he EDCF. QoS in EDCF EDCF com bines a collision based channel access ( CSMA/ CA) and priorit y packet scheduling t o priorit y CSMA/ CA ( see Figure 4.15) . Using t raffic cat egory classificat ion and priorit y CSMA/ CA, EDCF delivers different iat ed QoS services ( priorit ized QoS) .

Figu r e 4 .1 5 . ED CF QoS Ar ch it e ct u r e

4.2.2.2 Hybrid Coordination Function (HCF) HCF defines m ore rules t han EDCF allowing m ore cont rol of t ransm issions ( see Figure 4.16) . The HCF allocat es bandwidt h and t ransm ission opport unit ies ( TXOPs) using a hybrid coordinat or ( HC) t hat has t he highest access priorit y. HCF uses a cent ralized polling based approach sim ilar t o PCF. During CP, each TXOP st art s as defined by t he EDCF rules—t hat is, aft er AI FS plus backoff t im e, or when t he st at ion receives from t he HC a special poll fram e, referred as t he QoS Cont ent ion-

Free ( CF) - Poll. The HC sends t his QoS CF- Poll aft er it wait s for t he wireless m edia t o be idle for PI FS. The QoS CF- Poll specifies TXOP, which is t he t im e int erval during which t he st at ion has t he right t o t ransm it . During t he CFP, only t he polled st at ion is allowed t o t ransm it . The CFP ends aft er t he t im e announced in t he beacon fram e or by a CF- End fram e from t he HC.

Figu r e 4 .1 6 . H CF CFP Re pe t it ion I n t e r va l

I n order t o provide quant it at ive QoS services, HCF requires a signaling process t hat inform s t he HC about t he t ransm ission requirem ent s of each t raffic st ream at each st at ion. Using t his inform at ion, t he HC will det erm ine which st at ions need t o be polled, when, and which TXOP should be grant ed. The det ailed signaling process will be discussed in t he next sect ion. TXOP is grant ed per- st at ion—t hat is, t he HC does not specify which t raffic st ream should be t ransm it t ed on t he channel. I t is up t o t he st at ion t o select t he t raffic st ream t o be t ransm it t ed.

QoS in HCF As shown in Figure 4.17, HCF cont ains t he following QoS m echanism s: polling based channel access, QoS signaling, and t raffic st ream classificat ion. These m echanism s enable t he delivery of quant it at ive QoS services based on t he applicat ions' QoS requirem ent s ( TSPEC) . However, I EEE 802.11e does not define t he algorit hm s t o set HCF param et ers and it does not define t he relat ionship bet ween t hese param et ers and t he required QoS. Also, I EEE 802.11e does not present algorit hm s t o com put e t he polling sequence and t he t im e int ervals provided in t he HCF CFPoll.

Figu r e 4 .1 7 . H CF QoS Ar ch it e ct u r e

4.2.3 QoS Signaling The st andard defines t he QoS signaling which support s t he operat ion of t he QoS m echanism s. There are t wo form s of QoS signaling for t raffic originat ing at t he QoS st at ions: Queue St at e I ndicat or and Traffic Specificat ion. QoS st at ion t ransm it s t he m essage t hat cont ains Queue St at e or Traffic Specificat ion t o HC. QoS st at ion can t ransm it t he signaling m essage in t hree different ways. 1. Transm it during CFP int erval using HCF: QoS st at ion sends t he signaling m essage during t he CFP int erval t hrough t he HCF polling process. 2. Transm it during CP int erval using EDCF: The st at ion sends t he signaling m essage and cont ends wit h ot her st at ions for channel access. 3. Transm it during CP int erval using Cont rolled Cont ent ion ( CC) : The st at ion sends t he signaling m essage during Cont rolled Cont ent ion I nt erval ( CCI ) . The cont rolled cont ent ion process provides st at ions wit h t he opport unit y t o request TXOPs by sending resource request s, wit hout cont ending wit h ot her EDCF t raffic. Each cont rolled cont ent ion inst ance st art s when t he HC sends a specific cont rol fram e t hat forces legacy I EEE 802.11 st at ions t o set t heir NAV unt il t he end of t he cont rolled cont ent ion int erval, t hus t hey will refrain from t ransm ission during t he cont rolled cont ent ion int erval ( see Figure 4.18) .

Figu r e 4 .1 8 . N AV Se t t in g

4.2.3.1 Queue State Indicator Signaling A QoS st at ion t ransm it s t he queuing inform at ion of a t raffic st ream in t he st at ion t o t he HC. This will provide t he inform at ion t hat is helpful t o t he HC t o allocat e t he proper TXOP durat ion for t he st at ion t o m eet t he t raffic st ream 's dem and. The algorit hm t hat det erm ines t he TXOPs for each st at ion using t he queue inform at ion is not defined by t he st andard.

4.2.3.2 Traffic Specification Signaling This signaling support s t he resource reservat ion process in which a t raffic st ream in a QoS st at ion t ransm it s t he signaling m essage t hat includes t he t raffic specificat ion ( i.e, bandwidt h, delay requirem ent ) t o t he HC. The signaling cooperat es wit h t he resource reservat ion process at higher layers ( i.e., RSVP) . < Day Day Up >

< Day Day Up >

Chapter 5. HiperLAN Sect ion 5.1. I nt roduct ion Sect ion 5.2. Archit ect ure Sect ion 5.3. Physical Layer Sect ion 5.4. Dat a Link Cont rol ( DLC) Layer Sect ion 5.5. Convergence Layer Sect ion 5.6. QoS support < Day Day Up >

< Day Day Up >

5.1 Introduction The HiperLAN/ 2 set of st andards was int roduced by ETSI ( t he European Telecom m unicat ions St andards I nst it ut e) Proj ect BRAN ( Broadband Radio Access Net works) . ETSI worked in collaborat ion wit h a HiperLAN2 Global Forum , which is a consort ium of m ore t han 50 com panies est ablished t o m arket t his new st andard. I n addit ion, HiperLAN/ 2 was developed in harm ony wit h Japan's WLAN, referred t o as HiSWANa. HiSWANa was developed by Japan's ARI B ( Associat ion of Radio I ndust ries and Businesses) and it s Mult im edia Mobile Access Com m unicat ions ( MMAC) prom ot ion associat ion. Consequent ly, HiperLAN/ 2 and HiSWANa are very sim ilar. Relat ively few com panies have product s for HiperLAN/ 2 ( e.g., Panasonic, and Am phion) . HiperLAN/ 2 is designed t o provide access t o ext ernal I P, Et hernet , I EEE 1394, ATM, and 3G net works via an access point . HiperLAN/ 2 provides dat a rat es of up t o 54 Mbps using Ort hogonal Frequency Division Mult iplexing ( OFDM) radio t echnology for a range of up t o 150 m et ers. This generat ion of st andards support s bot h asynchronous dat a and applicat ions t hat require QoS support . Not ice t hat t his st andard, as ot her st andards, defines signaling procedures required for t he im plem ent at ion of such QoS support . However, t he algorit hm s t hat use t his signaling inform at ion are not defined and are left for t he developer. The Media Access Cont rol ( MAC) prot ocol im plem ent s a form of dynam ic t im e division duplex ( TDD) and dynam ic t im e division m ult iple access ( TDMA) t o provide connect ion orient ed service and QoS support . The st andard also support s st at ion m obilit y wit h speeds of up t o 10 m s. The predecessor of HiperLAN/ 2, HiperLAN/ 1, was geared for ad hoc net working and asynchronous dat a. I t applied t he carrier- sense m ult iple access wit h collision avoidance ( CSMA/ CA) m echanism t o resolve cont ent ion. HiperLAN/ 1 is considered a best effort syst em in spit e of t he fact t hat it included som e prelim inary m echanism s for QoS support . < Day Day Up >

< Day Day Up >

5.2 Architecture 5.2.1 Network Topology HiperLAN/ 2 dist inguishes bet ween a business environm ent and a resident ial environm ent . For a business environm ent t he st andard envisions a num ber of Access Point s ( APs) covering a cert ain area which m ay or m ay not overlap, depending on t he business needs ( Figure 5.1) . This net work t opology is called t he cellular access net work configurat ion. An AP provider int erconnect s all m obile t erm inals ( MTs) associat ed wit h it . All com m unicat ion goes t hrough t he AP.

Figu r e 5 .1 . H ipe r LAN / 2 Bu sin e ss N e t w or k ( ce llu la r a cce ss n e t w or k con figu r a t ion )

The resident ial environm ent is sim ilar t o t he business environm ent . I n addit ion, it can be operat ed in an ad hoc m anner in which t he MTs com m unicat e wit h each ot her direct ly ( see Figure 5.2) .

Figu r e 5 .2 . H ipe r LAN / 2 Re side n t ia l N e t w or k ( a d h oc n e t w or k con figu r a t ion )

The net work t opology for t he resident ial environm ent is called an ad hoc LAN configurat ion. An ad hoc LAN configurat ion also requires a st at ion called a cent ral cont roller ( CC) , which provides net work configurat ion cont rol t o all devices wit hin t he subnet . A CC in an ad hoc LAN configurat ion is sim ilar t o t he AP in t he cellular access net work configurat ion. Unlike t he dedicat ed hardware required for t he AP, a CC is dynam ically select ed from t he HiperLAN/ 2 m obile devices. When an MT which act s as a CC leaves t he net work, t he CC responsibilit y can be handed over t o anot her MT. HiperLAN/ 2 allows for m ult iple subnet s and t heir respect ive CCs in t he residence environm ent , sim ilar t o m ult iple cells and t heir respect ive APs in t he business environm ent . Such m ult iple subnet s can coexist since t hey operat e at different frequencies. HiperLAN/ 2 allows for t wo m odes of operat ion: Cent ralized m ode and Direct m ode. I n Cent ralized m ode all t raffic has t o pass t hrough t he AP t hat m anages t he access t o t he wireless m edia. This t raffic includes bot h t raffic bet ween m obile devices in t he net work and t raffic bet ween t he m obile device and t he out side net work. I n Direct m ode t he wireless m edia access is st ill m anaged by t he CC. However, t he t raffic bet ween t he m obile devices is exchanged direct ly wit hout going t hrough t he CC. Since t he CC can also be connect ed t o a core net work, it is required t hat t he CC will be able t o work in bot h Direct and Cent ralized m odes.

5.2.2 Protocol Stack As shown in Figure 5.3, HiperLAN/ 2 defines t he following t hree layers: t he Convergence layer, t he Dat a Link Cont rol ( DLC) wit h it s MAC and ot her funct ionalit y, and t he Physical layer.

Figu r e 5 .3 . H ipe r LAN / 2 Pr ot ocol St a ck

The convergence layer provides t he prot ocol int erface bet ween t he upper layer and t he DLC layer. The st andard envisions t hat HiperLAN/ 2 provides wireless access t o t he ext ernal or core net works such as I nt ernet Prot ocol ( I P) net works, Asynchronous Transfer Mode ( ATM) net works, 3G net works, and net works t hat use I EEE 1394 ( Firewire) prot ocols. The dat a t ransport funct ion of t he convergence layer provides m essage form at t ransform at ion ( i.e., segm ent at ion and reassem bly funct ion) bet ween t he higher layer and t he DLC layer. The dat a link layer cont ains t hree m ain ent it ies: t he Radio Link Cont rol ( RLC) , t he Error Cont rol ( EC) , and Media Access Cont rol ( MAC) . The user dat a plane funct ion of t he DLC receives t he dat a packet s from t he upper layer ( i.e., CL layer) , provides t he error cont rol m echanism , and delivers t he DLC dat a packet s t hrough t he MAC. The RLC which is t he cont rol plane funct ion of t he DLC provides t he Radio Resource Cont rol ( RRC) , Associat ion Cont rol Funct ion ( ACF) , and DLC Connect ion Cont rol ( DCC) . The HiperLAN/ 2 MAC of t he AP/ CC cont rols all t ransm issions over t he wireless m edia. This includes uplink t raffic from t he st at ion t o t he AP/ CC and downlink t raffic t o t he st at ion. The MAC uses t im e division duplex ( TDD) and dynam ic t im e division m ult iple access ( TDMA) . < Day Day Up >

< Day Day Up >

5.3 Physical Layer The physical layer support s several m odes of t ransm issions: 6, 9, 12, 18, 27, 36, and 54 Mbps. Each m ode em ploys a different radio m odulat ion t echnique as shown in Table 5.1.

Ta ble 5 .1 . H ipe r LAN / 2 Ra dio M odu la t ion Te ch n iqu e s M odu la t ion

Code Ra t e

Ph ysica l La ye r Bit Ra t e ( M bps)

BPSK

1/ 2

6

BPSK

3/ 4

9

QPSK

1/ 2

12

QPSK

3/ 4

18

16QAM

9/ 16

27

16QAM

3/ 4

36

64QAM ( opt ional)

3/ 4

54

HiperLAN/ 2 operat es at 5 GHz frequency spect rum as shown in Figure 5.4. I n t he U.S., 300 MHz bandwidt h in U- NI I band has been allocat ed, in Japan 100 MHz ( wit h sharing rule) bandwidt h has been allocat ed, and in Europe 455 MHz bandwidt h has been allocat ed in license exem pt band. The channel spacing is 20 MHz, which allows high bit rat es per channel but st ill has a reasonable num ber of channels in t he allocat ed spect rum ( e.g., 19 channels in Europe) .

Figu r e 5 .4 . H ipe r LAN / 2 Fr e qu e n cy Spe ct r u m Alloca t ion

HiperLAN/ 2 uses ort hogonal frequency division m ult iplexing ( OFDM) . The t ransm ission unit is a burst , which consist s of a pream ble part and a dat a part , where t he lat t er could originat e from each of t he t ransport channels wit hin t he DLC.

HiperLAN/ 2 defines Dynam ic Frequency Select ion ( DFS) , which allows sharing of t he wireless m edia am ong several HiperLAN/ 2 net works. The AP chooses frequencies based on a frequency select ion process considering int erference m easurem ent s. I n addit ion, t he AP/ CC will adapt it s t ransm ission m ode ( speed, code, and m odulat ion) based on it s m easurem ent s of t he t ransm ission link qualit y. Moreover, t he m obile st at ions, based on t heir m easurem ent s, m ay suggest t he preferred t ransm ission m ode. < Day Day Up >

< Day Day Up >

5.4 Data Link Control (DLC) Layer As m ent ioned earlier, t he DLC layer cont ains t hree sublayers: Error Cont rol ( EC) , Radio Link Cont rol ( RLC) , and Media Access Cont rol ( MAC) . I n addit ion, t here are logical and t ransport channels as shown in Figure 5.5. Logical channels are ident ified by t he t ype of m essage t hey carry while t ransport channels are ident ified by t he m essage form at and t he channel access m et hod. The cont rol m essages and user dat a m essages ( which originat e from t he DLC or higher layer) are m apped t o t he appropriat e logical channel based on t heir cont ent s. Then t he m essages are passed t o t he t ransport channel in order t o const ruct t he MAC fram e and receive t he appropriat e channel access. I n t he next subsect ions we int roduce t he funct ionalit y of each DLC sublayer, t he MAC fram e, and t he t ransport and logical channels. Then we describe MAC prot ocol exam ples.

Figu r e 5 .5 . D LC Pr ot ocol St a ck a n d M AC Fr a m e Con st r u ct ion

5.4.1 Radio Link Control (RLC) The RLC m anages t he net work and t hus exchanges cont rol dat a bet ween t he AP and t he MTs. The AP can have m ult iple RLC inst ances, where each RLC inst ance associat es wit h an MT. There is only one RLC inst ance in an MT.

The RLC has t hree funct ions: Radio Resource Cont rol ( RRC) , Associat ion Cont rol Funct ion ( ACF) , and DLC Connect ion Cont rol ( DCC) . The RRC is responsible for det ect ing and efficient ly using t he available radio resources. I t m anages handover, dynam ic frequency select ion, st at ion alive/ absent , power saving, and power cont rol. The RRC select s t he frequency range on which t he com m unicat ion will be conduct ed and, when needed, it decides t o m ove t o a different frequency range. This is done based on t he AP/ CC own channel m easurem ent s as well as t he m obile st at ion channel m easurem ent s. To preserve t he bat t ery power of t he st at ions, t he RRC defines when and for how long t o put t he MTs int o sleep m ode. The lengt hs of sleep int ervals are negot iat ed bet ween t he AP/ CC and t he MTs. The ACF m anages t he associat ion process for aut hent icat ion, key m anagem ent , associat ion, disassociat ion, and encrypt ion. A st at ion t hat want s t o com m unicat e over t he wireless m edia is required t o be associat ed wit h an AP/ CC. This is due t o t he fact t hat t he AP/ CC allocat es wireless resources for each associat ed st at ion. Moreover, t he MAC is cent rally cont rolled by t he AP/ CC in bot h Cent ralized Mode and Direct Mode. When a new MT j oins t he net work by perform ing t he associat ion process wit h t he AP's ACF, t he ACF assigns a unique MAC I dent ificat ion num ber ( MAC I D) t o t he MT. The ACF also m aint ains inform at ion of t he MT's capabilit ies in t erm s of whet her encrypt ion and aut hent icat ion are perform ed, and which encrypt ion and aut hent icat ion m et hods are em ployed. The DCC set s up, m aint ains, and releases user connect ions. I f any kind of QoS support is request ed, t he connect ion has t o supply t he QoS param et ers. The specificat ion of t hese param et ers is not part of t he HiperLAN/ 2 st andard. The connect ion set up includes procedures for cent ralized m ode or direct m ode, procedures for connect ion release, and procedures for j oining and leaving m ult icast groups.

5.4.2 Error Control (EC) The EC is responsible for det ect ion and recovery from t ransm ission errors on t he wireless m edia. I t also ensures in- sequence delivery of packet s. Each connect ion is support ed by an inst ance of t he EC. The EC is based on an Aut om at ic Repeat Request ( ARQ) algorit hm . Addit ional error correct ion t echniques can be em ployed. The ARQ schem e is based on a select ive repeat m echanism in which t he receiver request s t hat t he sender ret ransm it packet s det erm ined t o be erroneous. I n addit ion, t he EC m odule provides several m odes t o im prove t he t ransm ission reliabilit y: 1. Acknowledged m ode: The EC will ret ransm it t he acknowledgm ent s from t he receiver. Low lat ency can be m aint ained by a m echanism t hat discards packet s based on t im e inform at ion. 2. Repet it ion m ode: The EC will repeat t he t ransm ission. No acknowledgm ent s are available in t his m ode. Thus, t his m ode is t ypically used for broadcast dat a. 3. Unacknowledged m ode: This is for unreliable, low lat ency t ransm ission wit hout ret ransm issions. Therefore, unicast dat a can be sent using eit her t he Acknowledged m ode or t he Unacknowledged m ode. Broadcast dat a can be sent using eit her t he Repet it ion m ode or t he

Unacknowledged m ode. Mult icast dat a can be sent in t he Unacknowledged m ode or can be piggybacked int o ot her unicast t ransm issions.

5.4.3 Media Access Control (MAC) The HiperLAN/ 2 MAC t hat resides at t he AP/ CC cont rols all t he t ransm issions over t he wireless m edia. This includes 1) uplink t ransm issions from t he MTs t o t he AP/ CC, 2) downlink t ransm ission from t he AP/ CC t o t he MTs, and 3) direct t ransm ission am ong t he MTs. Direct t ransm ission is m andat ory for t he resident ial or ad hoc configurat ion. Therefore, t he MAC deploys t im e division duplex ( TDD) and dynam ic t im e division m ult iple access ( TDMA) . The channel is st ruct ured int o MAC fram es as shown in Figure 5.6. The MAC fram e has a fixed durat ion of 2 m s. Each MAC fram e st art s wit h t he Broadcast Channel ( BCH) durat ion.

Figu r e 5 .6 . M AC Fr a m e St r u ct u r e

As shown in Figure 5.7, t he fram e st ruct ure slight ly differs if t he AP/ CC has a sect ored ant enna. I n t his case, each phase is repeat ed in t im e, one for each sect or. The use of Direct Link ( DiL) wit h sect ored ant ennas is not specified.

Figu r e 5 .7 . M AC Fr a m e St r u ct u r e for Se ct or e d An t e n n a s

The MAC fram e consist s of: Broadcast Channel ( BCH) durat ion, Fram e Cont rol Channel ( FCH) durat ion, Access Feedback Channel ( ACH) durat ion and at least one Random Channel ( RCH) durat ion. I f t here is t ransm ission bet ween t he AP/ CC and t he MTs, t he Downlink ( DL) phase

and/ or Uplink ( UL) phase are included in t he MAC fram e. I f t here is t ransm ission am ong t he MTs ( direct m ode) , t he Direct Link ( DiL) phase is also included. The BCH durat ion is fixed. The durat ion of t he FCH, DL, DiL, and UL phases and t he num ber of RCHs are dynam ically det erm ined by t he AP/ CC. The BCH, FCH, ACH, and RCH cont ain cont rol m essages, whereas t he DL phase, DiL phase, and UL phase m ost ly cont ain user dat a and cert ain cont rol m essages. The st andard defines t he m essage form at t hrough t he t ransport and logical channels. There are six t ransport channels ( wit h t hree let t er abbreviat ions) : 1. Broadcast Channel ( BCH) 2. Fram e Channel ( FCH) 3. Access Feedback Channel ( ACH) 4. Long Transport Channel ( LCH) 5. Short Transport Channel ( SCH) 6. Random Access Channel ( RCH) There are t en logical channels ( wit h four let t er abbreviat ions) : 1. Broadcast Cont rol Channel ( BCCH) 2. Fram e Cont rol Channel ( FCCH) 3. Random Access Feedback Channel ( RFCH) 4. User Dat a Channel ( UDCH) 5. User Mult icast Channel ( UMCH) 6. User Broadcast Channel ( UBCH) 7. RLC Broadcast Channel ( RBCH) 8. Dedicat ed Cont rol Channel ( DCCH) 9. Link Cont rol Channel ( LCCH) 10. Associat ion Cont rol Channel ( ASCH) Figure 5.8 illust rat es t he MAC fram e st ruct ure, t he t ransport and logical channels and t heir relat ionships. All m essages are m apped t o specific logical channels based on t heir cont ent and t heir funct ionalit y. The m essages are const ruct ed according t o a form at defined by t he logical channel. Then t he m essages will pass t o t he t ransport channel t o receive t he appropriat e

t ransport services.

Figu r e 5 .8 . M AC Fr a m e St r u ct u r e , Tr a n spor t Ch a n n e l, a n d Logica l Ch a n n e l

Each MT's MAC is uniquely ident ified by a MAC I D which was assigned by t he AP's RLC during t he associat ion process. The HiperLAN/ 2 MAC is connect ion orient ed. Before an applicat ion can t ransm it , it requires t he MT t o est ablish a connect ion wit h t he AP's RLC. This RLC assigns t o t his connect ion a DLC Connect ion I D ( DLCC I D) . I n t he cent ralized m ode ( CM) , t he DLC User Connect ion I D ( DUC I D) , which is obt ained as a com binat ion bet ween t he MAC I D and t he DLCC I D, uniquely ident ifies t he connect ion in t he cell. I n dist ribut ion m ode ( DM) , t he definit ion of DUC I D is slight ly different . DUC I D, which ident ifies t he connect ion in DM m ode, is t he com binat ion of source MAC I D, dest inat ion MAC I D, and DLCC I D. DUC I D is also used for ident ifying t he cont rol channels. DUC I D is essent ial for t he AP t o allocat e bandwidt h t o a specific connect ion ( bot h user dat a and cont rol) .

5.4.4 Logical Channel A logical channel is viewed as a generic t erm for dist inct dat a pat hs. Logical channels are defined by t he t ype of inform at ion t hey carry.

5.4.4.1 Broadcast Control Channel (BCCH) The BCCH is used in t he downlink direct ion and carries BCCH inform at ion about t he ent ire wireless cell t o all t he st at ions wit hin t he cell. The BCCH m essage is only forwarded t o BCH t ransport channel and event ually is t ransm it t ed on t he BCH. BCCH cont ains: Net I D ( t he net work or t he cell ident ifier) , AP I D ( AP ident ifier) , AP TX level ( indicat ing t he t ransm ission

power of t he AP) , AP RX UL Level ( indicat ing t he expect ed recept ion power level of t he AP) . Moreover, t he BCCH also cont ains a point er or a locat ion and t he lengt h of t he FCH durat ion and RCH durat ion in t he MAC fram e as shown in Figure 5.9. The support provided by BCCH is m andat ory for APs and CCs. The m obile st at ions should be able t o int erpret t he BCCH.

Figu r e 5 .9 . Th e Poin t e r s in BCCH

5.4.4.2 Frame Control Channel (FCCH) The FCCH is used in t he downlink direct ion and carries inform at ion t hat describes t he st ruct ure of t he MAC fram e visible at t he air int erface ( see Figure 5.10) . FCCH is also called Resource Grant s ( RGs) . An RG, which is per connect ion basis ( DUC I D) , corresponds t o a num ber of t ransport channels ( t he num ber of LCHs or SCHs) , t he PHY m odes t o be used, and t he locat ion in t he fram e where t he recept ion/ t ransm ission will t ake place. The connect ion can be a cont rol connect ion ( RBCH, DCCH, LCCH) or a user dat a connect ion ( UDCH, UBCH, UMCH) . I n addit ion, FCCH includes announcem ent s of t he em pt y part s in t he MAC fram e. FCCH becom es an I nform at ion Elem ent ( I E) of t he FCH t ransport channel. The FCH can cont ain m ult iple I Es. Therefore, t he FCH's durat ion is variable. The support of t he FCCH is m andat ory for APs and CCs. MTs are required t o be able t o int erpret t he FCCH.

Figu r e 5 .1 0 . Fr a m e Con t r ol Ch a n n e l ( FCCH ) in FCH

5.4.4.3 Random Access Feedback Channel (RFCH) The RFCH carries inform at ion for st at ions t hat have used t he RCH in t he previous MAC fram e. This inform at ion cont ains t he result of t heir access at t em pt s. The RFCH m essage, which is t ransm it t ed once per MAC fram e, is included in t he ACH ( see Figure 5.11) . The support of t he RFCH is m andat ory for APs and CCs. MTs are required t o be able t o int erpret t he RFCH.

Figu r e 5 .1 1 . ACH Ope r a t ion

5.4.4.4 RLC Broadcast Channel (RBCH) The RBCH is used in t he downlink direct ion in Cent ralized Mode ( CM) and m ay be used by an MT when it com m unicat es wit hout t he AP/ CC in Direct Mode ( DM) . RBCH carries broadcast cont rol inform at ion about t he whole radio cell. RBCH t ransm ission is t riggered by t he AP/ CC in t he cent ralized m ode and is det erm ined by t he originat ing MT in t he direct m ode. The following inform at ion m ay be included in t he downlink RBCH: broadcast RLC m essages, t ransm ission of t he assigned MAC I D t o a non- associat ed m obile t erm inal, convergence layer I D inform at ion,

and encrypt ion relat ed inform at ion. Figure 5.12 shows t he RBCH fram e const ruct ion process. The support of t he downlink RBCH in AP/ CC and MTs is m andat ory.

Figu r e 5 .1 2 . RBCH Fr a m e Con st r u ct ion Pr oce ss

5.4.4.5 Dedicated Control Channel (DCCH) The DCCH is used t o carry RLC m essages. I t is used in t he uplink, downlink, and Direct Mode com m unicat ion. The DCCH m ay also be used for t he t ransm ission of Resource Request s ( RRs) in t he uplink direct ion. DCCH is t ransm it t ed t hrough LCH or SCH or RCH t ransport channel. Figure 5.13 shows t he DCCH fram e const ruct ion process. The support of t he DCCH in bot h MTs and AP/ CC is m andat ory. I t is required in Direct Mode if t his feat ure is support ed.

Figu r e 5 .1 3 . D CCH Fr a m e Con st r u ct ion Pr oce ss

5.4.4.6 User Broadcast Channel (UBCH) The UBCH is used t o t ransm it user broadcast dat a. The UBCH can be sent by t he AP/ CC in t he downlink and by an MT in t he direct link. The UBCH is t ransm it t ed in repet it ion or unacknowledged m ode. UBCH is t ransm it t ed t hrough LCH t ransport channel. UBCH fram e const ruct ion process is shown in Figure 5.14. The support of t he UBCH in bot h MTs and AP/ CC is m andat ory. I t is required in Direct Mode if t his feat ure is support ed.

Figu r e 5 .1 4 . UM CH / UBCH / UD CH Fr a m e s Con st r u ct ion Pr oce ss

5.4.4.7 User Multicast Channel (UMCH)

The UMCH is used t o t ransm it user m ult icast dat a. The UMCH can be sent by t he AP/ CC in t he downlink and by an MT in direct link. The UMCH is t ransm it t ed in repet it ion or unacknowledged m ode. UMCH is t ransm it t ed t hrough LCH t ransport channel. UMCH fram e const ruct ion process is shown in Figure 5.14. The support of t he UMCH in bot h MTs and AP/ CC is m andat ory. I t is required in Direct Mode if t his feat ure is support ed.

5.4.4.8 User Data Channel (UDCH) The UDCH is used t o t ransm it user dat a bet ween t he AP and t he MTs in Cent ralized Mode ( CM) or bet ween t wo MTs in Direct Mode ( DM) . A UDCH is always grant ed t oget her wit h zero or m ore t ransport channels for a connect ion t hat is announced in t he RG in t he FCCH. UDCH is t ransm it t ed t hrough LCH t ransport channel. UDCH fram e const ruct ion process is shown in Figure 5.14. The support of UDCH for uplink and downlink is m andat ory for bot h MTs and AP/ CC. The support of t he UDCH for Direct Mode is m andat ory if t his feat ure is support ed.

5.4.4.9 Link Control Channel (LCCH) The LCCH is used in bot h t he uplink and downlink t o t ransm it error cont rol inform at ion such as ARQ feedback. I t is used in eit her t he Cent ralized or Direct Mode. The LCCH is also used in t he uplink t ransm ission of t he RRs. LCCH is t ransm it t ed t hrough t he SCH or RCH t ransport channels. LCCH fram e const ruct ion process is shown in Figure 5.15. The support of t he LCCH for uplink and downlink is m andat ory for bot h MTs and AP/ CC. The support of t he LCCH for Direct Mode is m andat ory if t his feat ure is support ed.

Figu r e 5 .1 5 . LCCH Fr a m e Con st r u ct ion Pr oce ss

5.4.4.10 Association Control Channel (ASCH) The ASCH is only used in t he uplink and carries new associat ion and handover request

m essages. These m essages are sent only by t he MTs t hat are not associat ed t o an AP/ CC. ASCH is t ransm it t ed t hrough t he RCH t ransport channel. ASCH fram e const ruct ion process is shown in Figure 5.16. The support of t he ASCH is m andat ory for AP/ CC and MTs.

Figu r e 5 .1 6 . ASCH Fr a m e Con st r u ct ion Pr oce ss

5.4.5 Transport Channels The various logical channels are m apped int o corresponding t ransport channels t hat describe t he basic m essage form at . 1. Broadcast Channel ( BCH) : The BCH carries t he BCCH in t he downlink direct ion. I t s support is m andat ory for AP/ CC and st at ions. 2. Fram e Channel ( FCH) : The FCH carries t he FCCH. I t is broadcast ed in t he downlink direct ion and is m andat ory for all AP/ CC and st at ions. 3. Access Feedback Channel ( ACH) : The ACH carries RFCH in t he downlink. I t s support is m andat ory for AP/ CC and st at ions. 4. Long Transport Channel ( LCH) : The LCH carries user dat a for t he connect ions relat ed t o t he grant ed UDCHs, UBCHs, and UMCHs, as well as cont rol inform at ion for t he connect ions relat ed t o t he DCCH and RBCH. I t s support is m andat ory for AP/ CC and st at ions. 5. Short Transport Channel ( SCH) : The SCH carries short cont rol inform at ion for t he DCCH, LCCH, and RBCH. I t s support is m andat ory for AP/ CC and st at ions. 6. Random Access Channel ( RCH) : The RCH is defined for t he purpose of giving a st at ion

t he opport unit y t o send cont rol inform at ion t o t he AP/ CC when it has no grant ed SCH available. I t can carry RRs, ASCH, and DCCH dat a. I t s support is m andat ory for AP/ CC and st at ions.

5.4.6 Resource Request and Resource Grant The st andard defines t he resource request ( RR) and resource grant ( RG) signaling m echanism t o support t he bandwidt h allocat ion algorit hm , which plays an im port ant role in QoS support . The st andard does not define t he bandwidt h allocat ion algorit hm . Figure 5.17 shows a sim plified diagram of t he bandwidt h allocat ion input and out put .

Figu r e 5 .1 7 . Ba n dw idt h Alloca t ion

Because HiperLAN/ 2 MAC is connect ion orient ed, bandwidt h allocat ion is perform ed per connect ion basis. Each RR and RG serves for an individual connect ion. I n order for t he AP/ CC t o allocat e resources effect ively it needs t o know t he bandwidt h dem and of each connect ion represent ed by t he queue st at us ( i.e., queue lengt h) . Thus, t he st at ions will report t heir queue st at us in t he RR m essages t o t he AP/ CC. Using t he queue inform at ion received, t he QoS param et ers, and t he available net work resources, t he bandwidt h allocat ion algorit hm which resides at t he AP/ CC will det erm ine t he RG. The RG indicat es when a connect ion is allowed t o t ransm it and how m any packet s can be sent . The RG allocat ion algorit hm which is not defined by t he st andard needs t o be det erm ined by t he syst em developer. The RR can operat e in several ways: 1. Polling Resource Request ( Figure 5.18) : An MT init iat es t he RR by list ening t o t he poll. First , t he AP/ CC indicat es t he polling for an RR by set t ing t he RR Poll bit in t he FCCH m essage. Then, t he MT will t ransm it t he RR using LCCH wit h SCH t ransport channel on t he uplink.

Figu r e 5 .1 8 . Pollin g Re sou r ce Re qu e st

2. Unpolling Resource Request ( Figure 5.19) : An MT init iat es t he RR, which is sent t hrough DCCH wit h SCH t ransport channel ( collision- free t ransport ) and/ or LCCH wit h RCH t ransport channel ( collision- based t ransport ) . AP/ CC can cont rol t he access delay of t he RCH by changing t he num ber of cont ent ion slot s ( i.e., t he num ber of RCHs) . I f a collision occurs, t he st at ions are inform ed in t he ACH of t he next MAC fram e. Then, t he st at ion backs off a random num ber of access slot s.

Figu r e 5 .1 9 . Un pollin g Re sou r ce Re qu e st

< Day Day Up >

< Day Day Up >

5.5 Convergence Layer The convergence layer ( CL) m aps service request s from t he higher layers t o t he service offered by t he DLC. I t convert s higher layer packet s of fixed ( ATM) or variable ( Et hernet ) lengt h int o fixed- lengt h packet s, referred t o as Service Dat a Unit ( SDU) , t hat are used wit hin HiperLAN/ 2. For exam ple, CL will m ap t he 802.1p priorit y field int o a HiperLAN/ 2 priorit y class. < Day Day Up >

< Day Day Up >

5.6 QoS support HiperLAN/ 2 defines several QoS m echanism s t o provide QoS support for m ult im edia applicat ions. HiperLAN/ 2 m anages t he resources based on a connect ion- orient ed approach which enables t he finest granularit y of QoS support ( per flow and quant it at ive QoS services) . I n t his sect ion, we describe each QoS m echanism provided by HiperLAN/ 2.

5.6.1 Classification Each connect ion is t agged wit h DUC I D. I n cent ralized m ode ( CM) , DUC I D consist s of MAC I D and DLC connect ion I D ( DLCC I D) . I n dist ribut ed m ode ( DM) , DUC I D consist s of source MAC I D, dest inat ion MAC I D and DLCC I D. DUC I D enables classificat ion. Classificat ion ident ifies packet s based on DUC I D and forwards t hem t o t he appropriat e queue. Therefore, HiperLAN is capable of per- flow classificat ion which support s per- flow QoS services.

5.6.2 Channel Access As described in Sect ion 5.4.3, t he channel access is TDD/ TDMA, which is cont rolled by t he AP/ CC t hrough FCCH. Since TDD/ TDMA is a collision- free channel access schem e which provides t ight channel access cont rol, HiperLAN/ 2 can provide QoS support for applicat ions wit h st rict QoS requirem ent s.

5.6.3 Packet Scheduling The packet scheduling algorit hm 1) allocat es bandwidt h for connect ions in t erm s of t he num ber of packet s ( i.e., num ber of LCHs) and 2) det erm ines when a connect ion is allowed t o t ransm it . This inform at ion is defined in t he FCH I Es ( FCCH) . The packet scheduling algorit hm uses t he RR and RG, which are described in Sect ion 5.4.6. The st andard does not define t he packet scheduling algorit hm . I t j ust defines t he signalling m echanism such as RR and RG. Figure 5.20 sum m arizes t he HiperLAN/ 2 QoS archit ect ure which provides t he necessary m echanism s t o deliver per- flow quant it at ive QoS services.

Figu r e 5 .2 0 . H ipe r LAN / 2 QoS Ar ch it e ct u r e

< Day Day Up >

< Day Day Up >

Chapter 6. HomeRF Sect ion 6.1. I nt roduct ion Sect ion 6.2. Archit ect ure Sect ion 6.3. Physical Layer Sect ion 6.4. Media Access Cont rol ( MAC) Sect ion 6.5. QoS Support < Day Day Up >

< Day Day Up >

6.1 Introduction Hom eRF 2.0 st andard was developed by t he Hom eRF Working Group ( HRFWG) form ed by com panies such as Siem ens, Proxim , Nat ional Sem iconduct or, Xilinx, Mot orola, and AT&T. The group goal is t o est ablish t he m ass deploym ent of int eroperable wireless net working access devices t o bot h local cont ent and t he I nt ernet for voice, dat a, and st ream ing m edia in consum er environm ent s. Relat ive t o I EEE 802.11, few product s have been announced. Com panies such as Caym an Syst em s, Com paq, I nt el, Mot orola, and Proxim have announced Hom eRF 2.0 based com m unicat ion cards. Ut ilizing t he Hom eRF QoS support for voice, com panies such as Mot orola and Siem ens int roduced product s t hat include voice delivery. Mot orola int roduced t he Sim plefi syst em , which allows users t o t ransm it I nt ernet digit al audio t o t he exist ing st ereo equipm ent . Siem ens int roduced Voice Dat a Gat eway and Hom eRF based phones. Hom eRF 2.0 specificat ion defines a com m on int erface t hat support s wireless voice and dat a net working in t he hom e. The specificat ion allows 800 kbps rat e for isochronous voice and rat es of 1.6, 5, and 10 Mbps for asynchronous dat a t ransfer. Hom eRF specificat ion is backwards com pat ible wit h Hom eRF 1.2, which operat es at rat es of 0.8 Mbps and 1.6 Mbps. Hom eRF t arget s t o support a wide range of applicat ions by defining t he following t hree t ypes of dat a services: 1. Asynchronous dat a service: I t provides best effort service for asynchronous dat a applicat ions such as file t ransfer and em ail. 2. Priorit y asynchronous dat a service: I t provides priorit y service for asynchronous dat a applicat ions wit h loose QoS requirem ent s, such as st ream ing m edia ( video, audio) . 3. I sochronous dat a service: I t provides support for applicat ions wit h st rict QoS requirem ent s ( i.e., delay, delay j it t er) such as int eract ive voice and cordless phone. Hom eRF 1.x st andard did not describe roam ing procedures in t erm s of device hand- offs and resolut ion of m ult iple address conflict s. Hom eRF 2.0 added t he possibilit y of an ext ended net work t hat is an aggregat ion of individual wireless Hom eRF net works associat ed wit h t he sam e net work I D num ber. Such a net work is m anaged by roam ing capable Connect ion Point s ( CPs) . The st andard also defines roam ing procedures which include hand- offs, addresses, and dat a t ransfers bet ween roam ing devices. < Day Day Up >

< Day Day Up >

6.2 Architecture 6.2.1 Network Topology Before we describe t he Hom eRF net work t opology, we would like t o int roduce t he Hom eRF devices and t heir funct ionalit ies. The Hom eRF st andard cat egorizes t he Hom eRF devices int o t hree m ain t ypes, based on t he t ype of applicat ion t he devices support : 1. Asynchronous dat a device ( also called A- node) : These devices, such as PC, Lapt op, and PDA, support asynchronous dat a services ( i.e., file t ransfer, em ail) . 2. St ream ing dat a device ( also called S- node) : These devices, such as audio headset s, support priorit y asynchronous dat a services ( i.e., audio or video st ream ing) . 3. I sochronous dat a device ( also called I - node) : These devices, such as cordless phones, support isochronous dat a services ( i.e., t elephone call) . Som e devices can support m ult iple services. For exam ple, AI - nodes com bine t he funct ionalit ies of A- node and I - node. SI - nodes com bine t he funct ionalit ies of S- node and I - node. SA- nodes, such as I nt ernet appliances, com bine t he funct ionalit ies of S- node and A- node. The CP provides service m anagem ent for t he A- node, S- node and I - node. The CP has t he following funct ionalit ies: ●

●

●

A- node m anagem ent : I t support s A- nodes wit h power saving and provides dat a access bet ween t he A- node and t he I nt ernet or PCs on ot her net work segm ent s. S- node m anagem ent : I t provides session set up for t he S- node and assigns priorit y channel access t o t he S- node. I - node m anagem ent : I t provides connect ion set up for t he I - node, allocat es t he I - node dedicat ed net work resources ( i.e., bandwidt h) , and provides connect ivit y bet ween t he I node and t he Public Swit ch Telephone Net work ( PSTN) .

The Hom eRF st andard defines t wo m ain net work t opologies: Ad hoc Net work and Managed Net work. An ad- hoc net work ( see Figure 6.1) is a dist ribut ed net work ( sim ilar t o t he ad hoc net work concept in I EEE 802.11) in which Hom eRF devices can est ablish peer- t o- peer com m unicat ion. An ad hoc net work includes only A- nodes. There is no CP in an ad hoc net work. A- nodes in an ad hoc net work cannot operat e in t he power saving m ode.

Figu r e 6 .1 . Ad H oc N e t w or k

The Managed Net work ( see Figure 6.2) is a net work m anaged by t he CP. The Hom eRF devices can eit her est ablish peer- t o- peer com m unicat ion wit h each ot her or est ablish com m unicat ion t hrough t he CP depending on t he t ype of t he Hom eRF devices, for exam ple: ●

●

●

A- node: Peer- t o- peer com m unicat ion bet ween t he A- nodes, or com m unicat ion t o t he I nt ernet or PCs on ot her net work segm ent s t hrough t he CP. S- node: Peer- t o- peer dat a com m unicat ion bet ween t he S- nodes, or com m unicat ion wit h t he CP for signaling purposes ( i.e., session set up, priorit y assignm ent ) or for dat a t ransfer t o t he I nt ernet ( i.e., st ream ing video from t he I nt ernet ) . I - node: Com m unicat ion t o PSTN or PC t hrough t he CP for t ransm ission of bot h dat a and signaling. Peer- t o- peer com m unicat ion bet ween I - nodes is not allowed.

Figu r e 6 .2 . M a n a ge d N e t w or k

6.2.2 Protocol Stack As m ent ioned earlier, Hom eRF prot ocols use best effort asynchronous dat a service, priorit y asynchronous dat a service, and isochronous dat a service t o support sim ult aneous sessions for audio, video, and dat a. Bot h t he asynchronous and priorit y asynchronous dat a services are packet swit ched using a carrier sense m ult iple access ( CSMA/ CA) m edia access cont rol ( MAC) . The isochronous dat a service is for voice t hat requires st rict j it t er and lat ency and is support ed by a circuit swit ched, t im e division m ult iple access ( TDMA) MAC. Hom eRF support s TCP/ I P and User Dat agram Prot ocol ( UDP/ I P) t raffic as well as Digit al Enhanced Cordless Telecom m unicat ions ( DECT) t raffic ( see Figure 6.3) . DECT, which is a cordless phone st andard m ainly used in Europe, uses a TDMA based MAC ( i.e., assigns users wit h predet erm ined t im e slot s) . DECT provides speeds of up 2 Mbps. The im plem ent at ion of t he Hom eRF MAC differs from t he DECT MAC. However, because t he services it provides are sim ilar, t he DECT higher layers can be host ed on t he Hom eRF MAC wit h m inim al m odificat ion. The use of t he DECT prot ocol enables Hom eRF t o support call set up for t he isochronous dat a service and t o int eroperat e wit h t he PSTN.

Figu r e 6 .3 . H om e RF Pr ot ocol St a ck

< Day Day Up >

< Day Day Up >

6.3 Physical Layer The PHY layer is based on Frequency Hopping Spread Spect rum ( FHSS) and operat es in t he 2.4 GHz frequency band. I t provides 75 1 MHz channels for 1.6 Mbps dat a and voice com m unicat ion, and 15 5 MHz channels for 5 and 10 Mbps dat a com m unicat ion. Hom eRF specifies an adapt at ion m echanism t o ensure t hat t wo adj acent hops will not bot h be wit hin t he frequency range where int erference m ay occur. When int erference is det ect ed, t he hopping series set is exam ined t o find out if t wo consecut ive hops are bot h wit hin t he int erference range. I f such hops are found, t hen an at t em pt is m ade t o swit ch t hese hops wit h hops t hat are out side t he int erference range. For exam ple, if a m icrowave oven int erferes, t he hopping series will be adj ust ed t o avoid t he m icrowave frequency. < Day Day Up >

< Day Day Up >

6.4 Media Access Control (MAC) The st andard organizes t he com m unicat ion channel int o a superfram e st ruct ure t hat includes t wo periods—cont ent ion period and cont ent ion- free period. The durat ion of t he superfram e is fixed and has t wo possible values: 20 m s when t he superfram e cont ains only a cont ent ion period ( see Figure 6.4A) and 10 m s when t he superfram e cont ains bot h cont ent ion and cont ent ion- free periods ( see Figure 6.4B) . Each superfram e begins wit h a hopping period ( HOP) where Hom eRF devices change t heir hopping frequency. When t here is a cont ent ion- free period, t he superfram e cont ains t wo subfram es—cont ent ion- free period for voice t ransm issions and opt ional cont ent ion- free period for voice ret ransm issions ( see Figure 6.4B) .

Figu r e 6 .4 . H om e RF Su pe r fr a m e St r u ct u r e

6.4.1 Channel Access The st andard defines t wo channel access schem es for each period ( cont ent ion and cont ent ionfree) : 1. Cont ent ion period: The channel access is CSMA/ CA wit h added priorit y access and t im e reservat ion m echanism s. The priorit y access feat ure provides priorit y access t o t he st ream ing m edia t ransm ission over t he asynchronous dat a t ransm ission. Therefore, t his channel access is used by A- nodes and S- nodes. 2. Cont ent ion- free period: The channel access is TDMA wit h t he re- t ransm ission opt ion. Since t his channel access provides isochronous dat a t ransm ission service ( i.e., voice) , it is used by t he I - nodes ( e.g., cordless phones) . As shown in Figure 6.5, t he cont ent ion period includes asynchronous dat a t ransm ission where t he st ream ing m edia sessions receive priorit y. The cont ent ion- free period includes voice t ransm issions and opt ional voice ret ransm issions.

Figu r e 6 .5 . Ch a n n e l Acce ss

6.4.2 Contention-based CSMA/CA with Priority Access and Time Reservation Cont ent ion- based CSMA/ CA wit h priorit y access and t im e reservat ion support s t wo m odes of asynchronous dat a access: legacy non- priorit ized dat a access and priorit y asynchronous dat a access. I n case of non- priorit ized access, st at ions t hat have dat a t o t ransm it access t he wireless m edium in an uncoordinat ed way. Consequent ly, t his access has t he lowest access priorit y. I n t he priorit y access, t he st at ions access t he wireless channel based on a priorit y access value det erm ined by t he CP. S- nodes will be grant ed priorit y access whereas A- nodes will access t hrough t he non- priorit ized service. The CSMA/ CA m echanism support s a m ult i- rat e dat a service of 800 kbps, 1.6 Mbps, 5 Mbps, and 10 Mbps. Sim ilar t o I EEE 802.11 ( see Chapt er 4) , st at ions avoid collisions by list ening t o t ransm issions on t he wireless m edium . The st at ion will t ransm it only if it senses t hat t he m edium is free for a cert ain period of t im e referred t o as I nt erfram e Space ( I FS) . Then, t he st at ion random ly select s a t im e slot wit hin t he cont ent ion window t o st art t ransm ission. A collision is discovered by t he fact t hat t he sending st at ion does not receive t he acknowledgm ent from t he receiving st at ion. I n t his case, t he st at ion will double t he cont ent ion window size and ret ransm it dat a by reselect ing t he t im e slot from t he new cont ent ion window. The I FS is sim ilar t o I EEE 802.11: ●

●

SI FS ( short I FS) is t he short est t im e bet ween t wo consecut ive packet t ransm issions and used when t here is no need t o perform a Clear Channel Assessm ent ( CCA) procedure before init iat ing a t ransm ission. DI FS is t he short est t im e bet ween t he last bit of a sequence of a com plet e dat a t ransm issions and t he first bit of a new packet sequence. This int erval includes t he t im e t o recover from any prior t ransm ission, perform a CCA procedure, and st art a t ransm ission. DI FS definit ion in Hom eRF st andard is slight ly different from DI FS definit ion in I EEE 802.11. DI FS in Hom eRF includes t he random ized t im e slot durat ion of t he cont ent ion window, which is not included in I EEE 802.11.

I n addit ion, t his MAC support s t im e reservat ion which enables t he unint errupt ed t ransm ission of a sequence of m ult iple dat a packet s. St at ions t hat list en t o t he exchange of t im e reservat ion t ransm issions will not at t em pt t ransm issions. Hence, providing a cont ent ion- free m edium is for t he st at ions part icipat ing in t his t im e reservat ion exchange. Tim e Reservat ions are used t o signal t hat t he t im e reservat ion is m ade, canceled, or request ed. An opt ional Request - To- Send ( RTS) and Clear- To- Send ( CTS) packet exchange ( sim ilar t o I EEE 802.11) can be used for t he reservat ion process in case of hidden nodes. I n t he hidden node case, t he RTS- CTS exchange ensures t hat packet s at t he source vicinit y as well as t he dest inat ion vicinit y will defer for

t ransm ission during t he reserved period of t im e. The st andard also defines CSMA/ CA wit h priorit y access ( see Figure 6.6) , which can support up t o eight priorit ies. Priorit y access can be achieved t hrough reserved t im e slot s ( slot # 1 up t o slot # 8) in t he cont ent ion window. The CP assigns t hese slot s. Priorit y access is used by st ream ing m edia applicat ions. A st ream ing m edia applicat ion wit h st rict er QoS requirem ent s will be assigned a lower slot num ber which t ranslat es in a short er DI FS. Ot her asynchronous dat a sessions select t he slot num ber of t he cont ent ion window excluding t he reserved slot s. Therefore, t he st ream ing m edia wit h t he lowest assigned slot num ber will have t he highest priorit y access. Because t he channel access schem e uses t he CSMA/ CA m echanism , t he t im e delay is variable and cannot be guarant eed.

Figu r e 6 .6 . CSM A/ CA w it h Pr ior it y Acce ss

6.4.3 Reserved Time TDMA with Retransmission Option Reserved t im e TDMA wit h re- t ransm ission opt ion operat es in t he cont ent ion- free period and is available only in a CP m anaged wireless net work. The CP m anages t he I - nodes' access t o t he wireless m edium by sending special cont rol packet s, or beacons, t hat cont ain TDMA inform at ion. Because all I - nodes' dat a t ransm ission get s t hrough t he CP, t here are t wo direct ions of t ransm ission: uplink t ransm ission ( from I - node t o CP) and downlink t ransm ission ( from CP t o I - node) . This MAC m ode is connect ion orient ed—t hat is, each voice call originat ed at an I - node is required t o est ablish a connect ion wit h t he CP. The CP will allocat e a downlinkuplink pair of t im e slot s t o each call. Figure 6.7 shows a pot ent ial scenario of Hom eRF operat ion. Figure 6.7A shows t he channel allocat ion when only dat a are com m unicat ed using t he cont ent ion based prot ocol. The cycle is 20 m s long and is used by several dat a packet s and t heir respect ive acknowledgm ent s ( ACKs) ( Figure 6.7A) . I n Figure 6.7B, t wo voice conversat ions are added and t he cycle is reduced t o 10 m s. The Beacon, which is a m anagem ent cont rol packet , st art s each cycle and allocat es downlink and uplink TDMA t im e slot s. Aft er each cycle t he Frequency Hopping radio hops t o t he next frequency of t he frequency series.

Figu r e 6 .7 . Ex a m ple of H om e RF Ope r a t ion

The connect ion- orient ed TDMA prot ocol is referred t o as t he " single ret ransm ission aft er hop" ret ry prot ocol ( Figure 6.8) . This prot ocol support s high qualit y voice t ransm issions ( e.g., DECT based devices) t hat request bounded delays. The CP allocat es t im e slot s t o every act ive connect ion. This allocat ion m ay change if ot her connect ions are t erm inat ed. The TDMA acknowledgm ent is a sim ple piggyback acknowledgm ent insert ed in a lat er TDMA packet in t he reverse direct ion. I f t he CP does not receive a valid TDMA packet or if t he TDMA packet does not include an acknowledgm ent , t hen t he CP schedules a ret ry t im e slot for t he connect ion. This prot ocol provides bounded t im e delay because each t im e sensit ive connect ion is provided dedicat ed bandwidt h in bot h direct ions ( from / t o t he CP) ( i.e., dedicat ed TDMA slot s in every cycle) and dedicat ed bandwidt h in t he succeeding fram e if a ret ransm ission is required. A ret ransm ission, if needed, is done in t he consecut ive cycle which is less t han 10 m s lat er.

Figu r e 6 .8 . TD M A w it h Re t r a n sm ission

This prot ocol also provides support for a connect ionless broadcast service used by broadcast ing applicat ions whose packet s do not require acknowledgm ent s. Such applicat ions include I - node pages, cadence ringing, caller I D, and voice announcem ent . This service is provided by only using t he downlink t im e slot s. < Day Day Up >

< Day Day Up >

6.5 QoS Support Hom eRF st andard explicit ly defines t he QoS service it can support . As m ent ioned at t he beginning of t his chapt er, Hom eRF offers t hree m ain services: asynchronous dat a service, priorit y asynchronous dat a service, and isochronous dat a service. I n t his sect ion, we elaborat e t he QoS aspect s of each service. We will describe t he service charact erist ics and t he QoS m echanism s involved.

6.5.1 Isochronous Data Service This service is designed for applicat ions ( i.e., PSTN call, I nt ercom call, PC call, Conferencing) t hat originat e at an I - node which require st rict QoS cont rol in t erm s of bandwidt h, delay, and delay j it t er. To achieve t his service, t he following service charact erist ics and QoS m echanism s are involved: 1. Connect ion- orient ed service: An I - node sends t he connect ion request t o t he CP. The CP assigns t he connect ion I D and allocat es t he slot posit ion in t he cont ent ion- free period. This inform at ion is included in t he Beacon. The CP ident ifies t he connect ion by t he connect ion I D ( per- flow classificat ion) . 2. Delay and delay j it t er: I sochronous dat a service provides bounded delay and delay j it t er. Bounded delay and delay j it t er are achieved using a TDMA based channel access schem e wit h ret ransm ission opt ion. The CP allocat es a fixed t ransm ission durat ion for each connect ion. 3. Reliabilit y: The service allows t he ret ransm ission opt ion t o increase t he t ransm ission reliabilit y. 4. Priorit y: This service has t he highest priorit y. I n conclusion, t he isochronous dat a service can provide quant it at ive QoS services.

6.5.2 Priority Asynchronous Data Service This service, which is designed for applicat ions ( i.e., st ream ing video and audio) t hat originat e at S- nodes, involves t he following service charact erist ics and QoS m echanism s: 1. Session- orient ed service: An S- node sends t he connect ion request t o t he CP. The connect ion request cont ains a priorit y value ( 0- 7 value which is sim ilar t o t he I EEE 802.1D priorit y definit ion) or a delay/ delay j it t er requirem ent . The CP assigns t he session I D ( SI D) and det erm ines t he associat ed priorit y based on t he priorit y value or delay/ delay j it t er requirem ent inform at ion. The classificat ion is based on t he SI D. 2. Delay and delay j it t er: Delay and delay j it t er are not bounded but are less t han t he values provided by t he asynchronous dat a service. Priorit y CSMA/ CA channel access is used. Packet scheduling follows a st rict priorit y policy ( i.e., a session wit h a higher priorit y will receive service first ) .

3. Priorit y: This service has higher priorit y t han asynchronous dat a service but lower t han isochronous dat a service. From t he service charact erist ics present ed above, it can be concluded t hat priorit y asynchronous dat a service delivers qualit at ive ( relat ive) QoS service. The st andard does not define t raffic policing m echanism . A higher priorit y session can occupy and st arve t he bandwidt h of lower priorit y sessions and asynchronous dat a service.

6.5.3 Asynchronous Data Service This service is designed for applicat ions ( i.e., non- real- t im e applicat ions) t hat originat e at Anodes and do not have any specific QoS requirem ent s. This service has t he following service charact erist ics and QoS m echanism s: 1. Connect ionless service: This service does not require connect ion set up. 2. Delay and delay j it t er: Since t he channel access m echanism uses CSMA/ CA, t he delay and delay j it t er are variable. 3. Priorit y: This service has t he lowest priorit y. I n sum m ary, asynchronous dat a service delivers best effort services. < Day Day Up >

< Day Day Up >

Part 3: Wireless Metropolitan Area Networks

< Day Day Up >

< Day Day Up >

Chapter 7. IEEE 802.16 Sect ion 7.1. I nt roduct ion Sect ion 7.2. I EEE 802.16.1 Sect ion 7.3. Physical Layer Sect ion 7.4. Media Access Cont rol ( MAC) Sect ion 7.5. QoS Support Sect ion 7.6. I EEE 802.16a < Day Day Up >

< Day Day Up >

7.1 Introduction I EEE has been working on t he st andard for MAC and physical layer specificat ions for wireless m et ropolit an area net works ( WMANs or WirelessMAN™) . The goal of WMANs is t o provide highspeed wireless I nt ernet access sim ilar t o wired access t echnologies such as cable m odem , digit al subscriber line ( DSL) , Et hernet , and fiber opt ic. I EEE was m ot ivat ed by t he abilit y of t he wireless t echnologies t o cover large geographic areas wit hout t he need t o deploy wires. I EEE published t he first st andard in 2002. The first approved st andard, referred t o as I EEE 802.16.1, is geared t owards net work access bet ween buildings wit h ext erior ant ennas com m unicat ing wit h a cent ral radio base st at ion ( BS) in t he 10 t o 66 GHz frequency bands. I n parallel, t he European Telecom m unicat ions St andards I nst it ut e ( ETSI ) has been working on a st andard t hat serves sim ilar goals. This effort , referred t o as HiperAccess, was com plet ed in 2002, when ETSI decided t o adapt I EEE 802.16.1 as it s HiperAccess st andard. Since several WMANs can coexist and consequent ly int erfere wit h each ot her, t he I EEE organizat ion also published in 2001 a recom m ended pract ice, referred t o as I EEE 802.16.2 for Local and Met ropolit an Area Net works, such t hat m ut ual int erferences bet ween coexist ing WMANs are m inim ized. I EEE is st ill working on an addit ion t o t he 802.16.1 st andard t hat det ails MAC and physical layer specificat ions for t he 2 t o 11 GHz bands. This effort is referred t o as I EEE 802.16a. The European organizat ion, ETSI , has been working on a sim ilar proj ect , referred t o as HiperMAN. The organizat ions are working closely t oget her and it is expect ed t hat I EEE 802.16a will be adapt ed as HiperMAN. I EEE 802.16 is also working t o int roduce an am endm ent , referred t o as P802.16c, t hat will det ail Syst em Profiles for t he 10 t o 66 GHz band. These syst em profiles help in t he deploym ent of int eroperable syst em s. The I EEE 802.16 group realized t he need for m ult im edia applicat ions and t he required QoS support . Therefore, I EEE 802.16 has included a num ber of QoS signaling m echanism s. However, t he algorit hm s t hat use such signaling m echanism s in order t o provide QoS support are vendor specific and are left out of t he st andard. This allows vendors t o different iat e t heir product s but st ill be int eroperable. < Day Day Up >

< Day Day Up >

7.2 IEEE 802.16.1 I EEE 802.16 was designed considering t hat a m et ropolit an area can include hundreds or t housands of subscribers which require high- speed connect ions. The MAC is built t o accom m odat e a point t o m ult ipoint t opology, reflect ing t he fact t hat t he BS serves m ult iple subscriber st at ions ( SSs) . The 802.16a MAC will add t he capabilit y of serving m esh net works. The MAC addresses t he high- speed QoS requirem ent s wit h a flexible design of uplink ( SS t o BS) and downlink ( BS t o SS) channels. The BS has full cont rol on t he bandwidt h allocat ion on bot h channels. Access allocat ion is provided by t he BS via a request - grant m echanism , in which t he SS request s access and t he BS grant s access. The st andard realizes t hat users m ay have a variet y of needs result ing from t heir legacy voice syst em s, voice over I P syst em s, and t heir TCP/ I P packet ized syst em s. I t also realizes t hat , due t o different applicat ions, users m ay execut e applicat ions t hat result in eit her cont inuous or burst y t raffic, wit h different QoS requirem ent s.

7.2.1 Architecture 7.2.1.1 Network Topology The I EEE 802.16.1 is geared t oward point t o m ult ipoint ( PMP) com m unicat ion t opology, in which a BS com m unicat es wit h num erous st at ions, or SSs. All connect ions need t o go t hrough t he BS ( i.e., t here is no direct com m unicat ion bet ween t he SSs) . Due t o t he short wavelengt h of t he 10 t o 66 GHz frequency bands, wit h channels of 25 or 28 MHz, t he st andard requires line of sight bet ween t he BS and SS. Dat a rat es in excess of 120 Mbps can be support ed. The I EEE 802.16a ext ension effort support s significant ly reduced dat a rat es at 2 t o 11 GHz frequency bands. This approach does not require line of sight and support s m esh t opology in which connect ions do not need t o go t hrough t he BS. Figure 7.1 shows an exam ple of a net work t opology in which PMP and m esh t opologies are used t o cover a large m et ropolit an area. The in- building com m unicat ion can be provided t hrough wireless LANs ( e.g., I EEE 802.11) or via an im plem ent at ion of t he 802.16a ext ension. Hence, I EEE 802.16 can provide a com plet e wireless com m unicat ion solut ion for t he t arget ed m et ropolit an area.

Figu r e 7 .1 . Poin t t o M u lt ipoin t N e t w or k Topology

7.2.1.2 Protocol Stack The st andard is divided int o various layers as shown in Figure 7.2. The MAC layer is divided int o t hree sublayers: The Service Specific Convergence Sublayer ( CS) , t he MAC Com m on Part Sublayer ( CPS) , and t he Securit y Sublayer.

Figu r e 7 .2 . I EEE 8 0 2 .1 6 Pr ot ocol La ye r

The Service Specific Convergence Sublayer ( CS) t ransform s incom ing net work dat a, received t hrough t he CS service access point ( SAP) int o MAC dat a packet s, preserving or enabling QoS, and allowing bandwidt h allocat ion. This t ransform at ion m aps ext ernal net work inform at ion int o I EEE 802.16 MAC inform at ion, such as service flow and connect ion I D. The current st andard det ails t wo Convergence Sublayer specificat ions for ATM CS and for Packet CS. The Packet CS is defined for m apping ext ernal packet s such as I Pv4, I Pv6, Et hernet , and virt ual local area net work ( VLAN) . The MAC Com m on Part Sublayer ( CPS) provides access cont rol funct ionalit y, bandwidt h allocat ion, connect ion est ablishm ent , and connect ion m aint enance. The Securit y Sublayer provides securit y services including aut hent icat ion, key exchange, and encrypt ion. Dat a, physical layer ( PHY) cont rol, and m anagem ent inform at ion are exchanged bet ween t he MAC CPS and t he PHY via t he PHY SAP. The PHY layer includes m ult iple radio opt ions. < Day Day Up >

< Day Day Up >

7.3 Physical Layer The I EEE 802.16 defines t he physical layer for t he 10 t o 66 GHz frequency bands assum ing line of sight bet ween t he BS and SS. I t uses burst single- carrier m odulat ion in which all dat a are sequent ially t ransm it t ed in a single frequency. The st andard allows t he use of direct ional ant ennas t o increase capacit y and reduce int erference from adj acent t ransm issions. The physical layer support s wide channel bandwidt h of 20 or 25 MHz in t he U.S. and 28 MHz in Europe. Wit h 20 MHz channel widt h vendors can support up t o 96 Mbps, wit h 25 MHz up t o 120 Mbps, and wit h 28 MHz up t o 134 Mbps. The com m unicat ion pat h bet ween an SS and t he BS has t wo direct ions: uplink, from t he SS t o t he BS, and downlink, from t he BS t hat m ay reach m any SSs. Bot h uplink and downlink can operat e in different frequencies using Frequency Division Duplexing ( FDD) or share t he sam e frequencies using Tim e Division Duplexing ( TDD) . I n bot h FDD and TDD syst em s, t he uplink and downlink channels are st ruct ured int o fram es. I n TDD, t he fram e is divided int o t wo subfram es, uplink and downlink subfram es, where t he uplink subfram e follows t he downlink subfram e ( Figure 7.3) . I n FDD, t he uplink and downlink subfram es are concurrent in t im e but are carried on separat e frequencies ( Figure 7.4) . Each subfram e ( in TDD or FDD) consist s of a num ber of t im e slot s in which bot h TDD and FDD syst em s deploy TDM t ransm ission. All SSs and t he BS have t o be synchronized and t ransm it dat a burst s int o predet erm ined t im e slot s.

Figu r e 7 .3 . TD D Fr a m e St r u ct u r e

Figu r e 7 .4 . FD D Fr a m e St r u ct u r e

The st andard also support s full duplex subscriber st at ions ( which can t ransm it and receive sim ult aneously) and half duplex subscriber st at ions ( which can eit her t ransm it or receive) . For FDD syst em s, t he t im e slot assignm ent is different for full or half duplex subscriber st at ions. When t here are half duplex subscriber st at ions, t he BS cannot assign an uplink period for an SS which overlaps wit h t he SS downlink period. For TDD syst em s, due t o t he fact t hat t here is no overlapping period bet ween t he uplink and downlink channels, bot h full and half duplex subscriber st at ions can have t he sam e t im e slot s assignm ent .

7.3.1 Adaptive Data Burst Profiles The st andard support s adapt ive dat a burst profiling in which t ransm ission param et ers, such as m odulat ion and Forward Error Correct ion ( FEC) coding set t ings, can be m odified individually t o each SS on a fram e- by- fram e basis in bot h uplink and downlink t ransm issions. The st andard allows t hree t ypes of m odulat ion schem es: Quadrat ure Phase Shift Keying ( QPSK) , 16 Quadrat ure Am plit ude Modulat ion ( QAM) , and 64 QAM. Different com binat ions of m odulat ion and FEC provide different t ransm ission robust ness and t ransm ission speed ( i.e., 64 QAM delivers high t ransm ission speed but is prone t o int erference; whereas, QPSK delivers low t ransm ission speed but is t olerant t o int erference) . Dat a burst profiles, which include param et ers such as radio m odulat ion t ype and FEC, are ident ified by a code called I nt erval Usage Code ( I UC) . There are t wo t ypes of I UC: 1) Downlink I nt erval Usage Code ( DI UC) , which ident ifies t he downlink dat a burst profiles, and 2) Uplink I nt erval Usage Code ( UI UC) , which ident ifies t he uplink dat a burst profiles. DI UC is included in a MAC m essage called Downlink Channel Descript or ( DCD) m essage; whereas, UI UC is included in t he Uplink Channel Descript or ( UCD) m essage. DCD and UCD m essages are t ransm it t ed periodically by t he base st at ion in order t o define t he downlink and uplink channel charact erist ics, respect ively. < Day Day Up >

< Day Day Up >

7.4 Media Access Control (MAC) The 802.16.1 MAC is connect ion orient ed. All t raffic including inherent ly connect ionless t raffic is m apped int o a connect ion. This provides t he abilit y t o request bandwidt h wit h it s associat ed QoS and t raffic param et ers for every connect ion. Connect ions are ident ified by Connect ion I dent ifiers ( CI Ds) . The st andard allows bandwidt h provision t o t hese connect ions t o be eit her cont inuous or on dem and. Each connect ion is associat ed wit h a service flow ( SF) t hat is a unidirect ional flow of MAC packet s on a connect ion t hat is provided wit h a part icular QoS. The service flow defines t he QoS param et ers for t he packet s t hat are t ransm it t ed on t his connect ion. The Service Flow concept is cent ral t o t he 802.16 MAC since it provides t he m echanism for QoS and bandwidt h allocat ion process. The MAC has also som e reserved CI Ds for purposes such as m anagem ent , announcem ent s, and broadcast ing. Est ablished connect ions require act ive m aint enance depending upon t he t ype of service. For exam ple, unchannelized T1 services require virt ually no connect ion m aint enance since t hey have a const ant bandwidt h allocat ion. Channelized T1 services require som e m aint enance due t o t he dynam ic bandwidt h requirem ent s. Connect ions m ay be also t erm inat ed or adj ust ed based upon users' cont ract s. The st andard support s all of t hese connect ion m anagem ent funct ions t hrough t he use of st at ic configurat ion and dynam ic addit ion, m odificat ion, and delet ion of connect ions.

7.4.1 Channel Access As described earlier, t he com m unicat ion pat h has t wo direct ions—uplink and downlink. I n uplink t ransm issions ( SSs t o BS) , several SSs share t he channel in a TDMA fashion. The Uplink Map Message ( UL- MAP) is used t o provide t he channel access assignm ent t o t he subscriber st at ions. The UL- MAP which is t ransm it t ed by t he base st at ion at t he beginning of t he fram e defines t he uplink channel access as well as t he uplink dat a burst profiles ( i.e., UI UC) in t he current uplink subfram e ( see Figure 7.5B- C) . The SSs are allowed t o t ransm it dat a burst s at predet erm ined t im e slot s as indicat ed in t he UL- MAP. I n downlink t ransm issions ( BS t o SSs) , due t o t he fact t hat t here is only one st at ion ( BS) t ransm it t ing t he dat a, t he channel access is rat her sim ple. The dat a packet s are t ransm it t ed by t he BS t o all SSs and an SS only picks up t he packet s dest ined t o it . The Downlink Map Message ( DL- MAP) is used t o define t he downlink dat a burst profiles ( i.e., DI UC) in each t im e period in t he current downlink subfram e, in bot h TDD and FDD syst em s ( see Figure 7.5A) .

Figu r e 7 .5 . M AC Con t r ol I n for m a t ion

Bot h UL- MAP and DL- MAP indicat e t he st art ing t im e slot of each dat a burst ( see Figure 7.6) .

Figu r e 7 .6 . D L- M AP a n d UL- M AP

7.4.1.1 Downlink Subframe The st andard specifies different downlink fram es for FDD and TDD syst em s. The FDD downlink subfram e ( Figure 7.7) st art s wit h a pream ble, DL- MAP, and UL- MAP. The pream ble helps in t he physical layer t ransit ion and synchronizat ion. The DL- MAP defines burst s' st art t im es on t he downlink for bot h TDM and t im e division m ult iple access ( TDMA) m et hods. The downlink subfram e is divided int o TDM and TDMA port ions, where t he TDMA port ions follow t he TDM port ions. The TDM port ions are not separat ed by gaps or pream bles since t here is only one t ransm it t er, t he BS. The TDM port ions cont ain dat a t ransm it t ed t o one or m ore of t he following: full duplex SSs, half duplex SSs scheduled t o t ransm it lat er in t he fram e, and half duplex SSs not scheduled t o t ransm it in t his fram e. On t he ot her hand, t he TDMA port ions are separat ed by pream bles and gaps. The TDMA port ion is used t o t ransm it dat a t o any half duplex SSs scheduled t o t ransm it earlier in t he fram e t han t hey receive dat a. This allows an individual SS t o decode a specific port ion of t he downlink wit hout t he need t o decode t he ent ire downlink subfram e.

Figu r e 7 .7 . FD D D ow n lin k Su bfr a m e St r u ct u r e

The TDD downlink subfram e ( Figure 7.8) st art s wit h a pream ble, DL- MAP, and UL- MAP. The

pream ble helps in t he physical layer t ransit ion and synchronizat ion. The TDD downlink subfram e cont ains dat a packet s t hat are t ransm it t ed t o SSs and ends wit h a Transm it / Receive ( Tx/ Rx) t im e gap. I n t he TDD downlink subfram e t here are only TDM port ions.

Figu r e 7 .8 . TD D D ow n lin k Su bfr a m e St r u ct u r e

7.4.1.2 Uplink Subframe The uplink subfram e ( Figure 7.9) is used by t he SSs t o t ransm it t o t he BS.

Figu r e 7 .9 . Uplin k Su bfr a m e St r u ct u r e

The uplink subfram e includes t he following t hree periods: I nit ial Maint enance period, Request Cont ent ion Opport unit ies period, and Scheduled Dat a grant s period. The different periods are ident ified by t heir respect ive UI UC t hat is specific t o an uplink period. The BS announces t hese

periods and associat ed burst classes in t he preceding downlink subfram e's UL- MAP. The BS can specify such periods in any order and lengt h. The BS can group t he periods allocat ed t o t he I nit ial Maint enance and Request Cont ent ion Opport unit ies and leave t he rem aining periods for dat a t ransm ission. We provide next a m ore det ailed descript ion of each period. I n t he I nit ial Maint enance period, ident ified by UI UC = 2, t he SS sends t ransm issions or burst s needed t o carry out I nit ial Maint enance funct ions. For exam ple, ranging request s are sent by t he SS at t he init ializat ion phase and periodically t hereaft er. The BS uses such request s t o det erm ine net work delay and request power or downlink burst profile changes. I n t his period also new st at ions m ay j oin t he net work. Since several SSs can access t he channel sim ult aneously, collisions can occur in t his period. I n t he Request Cont ent ion Opport unit ies period, ident ified by UI UC = 1, t he SSs request bandwidt h based upon m ult icast and broadcast polls by t he BS. Since several SSs can access t he channel sim ult aneously, collisions can occur in t his period. I n t he Scheduled Dat a grant s period, ident ified by UI UC t hat defers from 1 or 2, t he SS t ransm it s dat a based on grant s allocat ed by t he BS. An SS t ransit ion gap separat es each Scheduled Dat a period t o allow t ransit ion from one period wit h a specific burst profile t o anot her period which m ay have a different burst profile. Each such period st art s wit h a pream ble t o allow t he new SS t o synchronize.

7.4.1.3 Contention Resolution Since several SSs can access t he channel sim ult aneously during t he I nit ial Maint enance and Request periods, t heir respect ive t ransm issions can collide. I f such a collision occurs t he t ransm it t ing SSs will go t hrough a cont ent ion resolut ion process. I t is possible t hat t he SS will have m ult iple uplink service flows, each associat ed wit h it s CI D. The SS will consider t hese CI Ds or t heir QoS needs in t his process. The m andat ory process by t he st andard is based on a t runcat ed binary exponent ial backoff, wit h an init ial backoff window and a m axim um backoff window cont rolled by t he BS. The algorit hm is defined in t he st andard.

7.4.2 Bandwidth Request Mechanisms and Bandwidth Allocation The st andard defines various m echanism s for t he SS t o access t he shared uplink and request t ransm ission opport unit ies ( bandwidt h) and for t he BS t o grant such t ransm ission opport unit ies. The key m echanism s of t he request - grant process are t he bandwidt h request m echanism and t he bandwidt h allocat ion. These m echanism s enable m ult iple t ypes of service flows which support a wide range of applicat ions. First , we int roduce t he bandwidt h request and t he bandwidt h allocat ion m echanism s and t hen we describe each t ype of service flow offered by t he st andard.

7.4.2.1 Bandwidth Request Mechanism Bandwidt h is always request ed on a connect ion ( i.e., CI D) basis. Each connect ion in an SS request s bandwidt h t hrough an MAC m essage called BW Request m essage. BW Request can be sent in a st and- alone packet or piggyback on anot her packet . The request is based on t he num ber of byt es needed t o carry t he MAC packet wit hout considering t he physical layer overhead. The request ed bandwidt h can be eit her increm ent al, m eaning how m uch addit ional bandwidt h is needed, or aggregat e, m eaning how m uch t ot al bandwidt h is needed. When t he BS receives an increm ent al bandwidt h Request , it adds t he quant it y of bandwidt h request ed t o it s current allocat ed bandwidt h. When t he BS receives an aggregat e bandwidt h Request , it

replaces t he allocat ed bandwidt h wit h t he quant it y of bandwidt h request ed. Wit h piggyback BW Request , t he connect ion can only request increm ent al bandwidt h. Wit h st and- alone BW Request , t he connect ion can request eit her increm ent al or aggregat e bandwidt h. BW Request can be init iat ed direct ly by t he connect ion or init iat ed in response t o receiving a Polling Message from BS. BW Request can be t ransm it t ed during t he uplink subfram e in t wo different ways: 1) Cont ent ion request opport unit ies in which BW Request is t ransm it t ed in Request Cont ent ion period and 2) Cont ent ion- Free Request Opport unit ies in which BW Request is t ransm it t ed in t he predefined t im e slot of uplink subfram e indicat ed by I nform at ion Elem ent ( I E) of UL- MAP or is piggybacked wit h ot her packet s. The polling m echanism is geared for det ailed and flexible request s but requires m ore overhead. BS issues unicast polling packet s for t he SS as well as m ult icast and broadcast polls. The polling process does not include an explicit packet from t he BS t o t he SS. Rat her, t he SS is allocat ed bandwidt h for it s pot ent ial request in t he downlink subfram e field t hat det ails t he com ing uplink t ransm ission opport unit ies, referred t o as uplink m ap. The st andard also allows t he BS t o poll st at ions in groups or as a whole. This m ode can be used in cases where t he BS decides t hat t here is not enough bandwidt h t o support individual unicast polls or in cases where t here are m any inact ive SSs. Cert ain connect ion CI Ds are reserved for such m ult icast and broadcast polls. The SSs addressed in t his group polling will need t o cont end during t he uplink period t o send t heir t ransm ission request s.

7.4.2.2 Bandwidth Allocation As m ent ioned in t he previous sect ion, each SS request s bandwidt h on a connect ion basis. The BS grant s or allocat es bandwidt h by deploying one of t he following t wo m odes: Grant Per Subscriber St at ion ( GPSS) and Grant Per Connect ion ( GPC) . I n GPC m ode, t he BS allocat es bandwidt h t o individual connect ions ( CI D) . I n GPSS m ode, t he BS allocat es bandwidt h for individual SSs. I n t his case, t he SS is responsible for dist ribut ing t his allocat ed bandwidt h t o each connect ion wit hin t he SS. The SS decides which t raffic should be t ransm it t ed first , considering each applicat ion's t raffic needs and QoS requirem ent s. Thus, GPSS m ode requires m ore int elligence in t he SS t han does t he GPC m ode. This allows vendors t o design different SS configurat ions. A vendor m ay choose a GPC design for sim plicit y. I f t he 10 t o 66 GHz frequencies are used, t he st andard allows only SSs wit h GPSS m ode. The BS allocat es bandwidt h t o a connect ion or a st at ion by grant ing t he t ransm ission opport unit y in t he UL- MAP. The bandwidt h allocat ion is m ainly based on t he following fact ors: ● ●

●

The am ount of bandwidt h request ed by t he connect ions QoS param et ers of delay and bandwidt h needed by t he current applicat ions at t he SS as reflect ed in t he service flow param et ers Available net work resources

The BS can grant t ransm ission opport unit ies for m ult iple connect ions t o a single SS, each wit h a different QoS level. The BS m ay not grant bandwidt h t o t he request ing SS due t o reasons such as: t he bandwidt h request ed by t he SS is not available, t he BS received a request t hat included errors, and t he request did not go t hrough due t o collisions wit h ot her SSs or due t o errors in t he wireless m edia.

7.4.3 Service Flow Scheduling Service The st andard defines four t ypes of service flows t hat provide QoS support for a wide range of

applicat ions. The services include: Unsolicit ed Grant Service ( UGS) , Real- Tim e Polling Service ( rt PS) , Non- Real- Tim e Polling Service ( nrt PS) , and Best Effort ( BE) Service.

7.4.3.1 Unsolicited Grant Service (UGS) The Unsolicit ed Grant Service ( UGS) support s real- t im e service flows t hat generat e fixed size periodic dat a packet s, such as T1. This service provides periodic, fixed size grant s t hat avoid t he overhead and lat ency of frequent SS redundant request s. I t also ensures t hat grant s are available t o m eet t he cont inuous needs of t he service flow. A connect ion wit h UGS service flow is prohibit ed from using any cont ent ion request opport unit ies. The BS will not provide any unicast request opport unit ies for t his connect ion. The I nform at ion Elem ent ( I E) of t he UGS service flow includes param et ers such as t he UGS size, nom inal grant int erval, t he t olerat ed grant j it t er, and t he request / t ransm ission policy. The SS updat es t he BS regarding t he st at e of t he UGS service flow, using special fields called Slip I ndicat or ( SI ) flag in t he packet header. The SI flag indicat es excessive queue lengt h due t o condit ions such as a lost allocat ion m ap or due t o clock m ism at ch bet ween t he SS and t he BS. I n such cases t he BS m ay grant addit ional bandwidt h t o com pensat e for bandwidt h insufficiency. An SS t hat is grant ed UGS opport unit y cannot " st eal" bandwidt h. I n ot her words, t he SS cannot use a port ion of t he allocat ed bandwidt h t o send anot her bandwidt h request rat her t han sending dat a. Also it can not piggyback a request in t he t ransm it t ed dat a. However, t he SS can indicat e t hat it needs a poll for non- UGS service by set t ing a special field, referred t o as Poll- Me bit , in t he packet header. This m echanism saves bandwidt h, as t he BS can poll such st at ions only if t he Poll- Me bit is set . Once t he BS receives t his indicat ion, it m ay follow wit h individual polling such as rt PS and nrt PS.

7.4.3.2 Real-Time Polling Service (rtPS) The BS polls SSs wit h t he purpose of allocat ing t o t hese SSs bandwidt h specifically for t he purpose of m aking bandwidt h request s. The rt PS support s real- t im e service flows t hat have periodic dat a packet s of various sizes, such as MPEG st ream s. This service provides periodic, unicast request opport unit ies, t o respond t o t he needs of t he service flows. I t allows t he SS t o dynam ically specify t he size of t he request ed grant . Com pared t o UGS, t his service has addit ional overhead due t o t he polling process. However, it can handle variable grant sizes com pared t o UGS t hat can handle only fixed grant sizes. The st andard specifies t hat t he BS needs t o poll t he SS periodically in order t o allow it t o send unicast request opport unit ies. The polling frequency ( t he num ber of polls in a cert ain period of t im e) is not specified in t he st andard. I n com m on pract ice, t he BS should poll t he SS frequent ly enough t o m eet t he delay and bandwidt h requirem ent s of real- t im e applicat ions. I n t his case, t he SS is prohibit ed from using any cont ent ion request opport unit ies in order t o avoid unpredict able delays. A connect ion wit h rt PS service flow is allowed t o " st eal" bandwidt h if it is in t he GPSS m ode. rt PS service flow is also allowed t o piggyback a BW Request on a dat a packet .

7.4.3.3 Non-Real-Time Polling Service (nrtPS) The Non- Real- Tim e Polling Service ( nrt PS) support s non- real- t im e service flows t hat have dat a packet s of various sizes, such as FTP. The service provides consist ent unicast request opport unit ies even during net work congest ion. The st andard specifies t hat a BS needs t o poll t he SS on a regular basis ( periodically or nonperiodically) and allow it t o send unicast request opport unit ies. I n t his case, t he SS is allowed t o use cont ent ion request opport unit ies for t he connect ion. I n ot her words, t he SS can sim ult aneously send unicast request packet s in response t o t he BS Poll and cont ent for request opport unit ies. A connect ion wit h nrt PS service flow is allowed t o " st eal" bandwidt h if it is in GPSS m ode. nrt PS service flow is allowed t o

piggyback a BW Request on a dat a packet .

7.4.3.4 Best Effort (BE) Service The Best Effort ( BE) Service support s service flows t hat do not require QoS support . The SS issues it s request s in a cont ent ion period. An SS t hat is grant ed a request via BE is allowed t o " st eal" bandwidt h if it is in GPSS m ode. BE Service flow is allowed t o piggyback a BW Request on a dat a packet . The st andard det ails ext ensive signaling t echniques and access m echanism s for each service flow, but t he det ails of bandwidt h allocat ion, scheduling, and reservat ion m anagem ent int elligence are out of t he st andard's scope and are left t o be vendor specific. The abilit y t o em ploy different com binat ions of t hese access m echanism s allows vendors t o different iat e t heir product s, t ailor solut ions t o unique needs and users, opt im ize syst em perform ance, and use different bandwidt h allocat ion algorit hm s while m aint aining consist ent int eroperabilit y. For exam ple, cont ent ion m ay be act ivat ed t o avoid polling of SSs t hat have been inact ive for a long period of t im e.

7.4.4 Network Entry and Initialization An SS wishing t o ent er t he net work has t o go t hrough t he following st eps: 1. Scan for a downlink channel and est ablish synchronizat ion wit h t he BS. The SS can eit her recall t he downlink channel from it s m em ory or cont inuously scan possible channels of t he downlink frequency band unt il it finds a valid downlink signal. Aft er it finds a downlink channel it needs t o synchronize and at t em pt t o acquire t he channel cont rol param et ers for t he downlink by searching for t he DL- MAP sent by t he BS. 2. Obt ain t ransm it param et ers. The SS searches for an uplink Channel Descript or m essage from t he BS in order t o ret rieve t he t ransm ission param et ers for a possible uplink channel. 3. Adj ust local param et ers ( e.g., t ransm it power) based on approved values or m essages from t he BS. 4. Negot iat e basic capabilit ies. The SS inform s t he BS of it s basic capabilit ies by sending an appropriat e m essage t o t he BS. The BS acknowledges t his m essage. 5. Aut horize SS and perform key exchange. The BS goes t hrough a process of aut horizing t he SS t o ent er t he net work and exchanges securit y keys wit h t he SS. 6. Perform regist rat ion. The BS sends addit ional m anagem ent m essages and t he SS becom es m anaged by t he BS. 7. Est ablish I P connect ivit y. The SS receives an I P address from t he BS. 8. Est ablish t im e of day. The SS and BS need t o have equal t im ing inform at ion. 9. Transfer operat ional param et ers. The BS sends addit ional configurat ion inform at ion t o t he SS.

< Day Day Up >

< Day Day Up >

7.5 QoS Support The principal m echanism of I EEE 802.16 st andard for providing QoS support is t o associat e a packet wit h a service flow. A service flow is a unidirect ional flow of packet s t hat provides a part icular QoS. The st andard det ails t he m echanism s of how t o allocat e bandwidt h and how t o send t he BW Request s in each service flow as described in t he previous sect ion. To sum m arize, t he service flows and t he support ed applicat ions are list ed below: ●

●

●

●

UGS Service flow support s real- t im e applicat ions wit h const ant bit rat e ( CBR) such as voice over I P and circuit em ulat ion. rt PS Service flow support s real- t im e applicat ions wit h variable bit rat e ( VBR) such as st ream ing video and audio. nrt PS Service flow support s non- real- t im e applicat ions which require bet t er service t han BE, such as high- bandwidt h FTP. BE Service flow support s applicat ions t hat do not have any QoS requirem ent s.

Each net work applicat ion, first , has t o regist er wit h t he net work. The net work will associat e t he applicat ion wit h a service flow by assigning an unique Service Flow I D ( SFI D) . All packet s m ust be t agged wit h t his assigned SFI D in order for t he net work t o provide t he appropriat e QoS. When t he applicat ion want s t o send dat a packet s, it is required t o est ablish a connect ion wit h t he net work and receives a unique CI D assigned by t he net work. Therefore, t he I EEE 802.16 dat a packet s include bot h CI D and SFI D. Next we describe QoS support from t he following t wo aspect s: QoS provision and QoS m echanism s.

7.5.1 QoS Provision I EEE 802.16 broadband wireless access st andard aim s t o provide fixed wireless access bet ween t he subscriber st at ion ( resident ial or business cust om ers) and t he I nt ernet Service Provider ( I SP) t hrough t he BS. The service m odel involves t he following st eps: ●

●

●

●

I n order t o receive net work service, t he cust om ers have t o subscribe or regist er wit h t he I SP or anot her ent it y t hat has t he aut horit y t o cont rol t he net work. Whenever t he cust om ers want t o use t he service ( i.e., t ransm it dat a) , t hey need t o est ablish t he connect ion wit h t he net work. I f cust om ers require QoS support for t heir applicat ions, t hey are required t o include a QoS param et er set t hat can be defined in any of t he following ways: explicit ly specifying all t raffic param et ers, indirect ly referring t o a set of t raffic param et ers via specifying a Service Class Nam e, or specifying a Service Class Nam e along wit h m odifying t he param et ers. Considering t his QoS param et er set and t he available net work resources, t he net work decides if it can provide t he appropriat e QoS support .

The st andard's set of t ools t hat support QoS for bot h uplink and downlink t raffic include: configurat ion and regist rat ion funct ions for service flows, a signaling funct ion for dynam ically est ablishing QoS based service flows and t raffic param et ers, scheduling and QoS t raffic param et ers for uplink and downlink service flows, grouping of service flow propert ies int o service classes t o allow grouping of request s.

The st andard's QoS provision, based on t he " envelopes" m odel shown in Figure 7.10, defines several set s of QoS param et ers: ●

●

●

ProvisionedQoSParam Set : A set of ext ernal QoS param et ers provided t o t he MAC, for exam ple, by t he net work m anagem ent syst em . Adm it t edQoSParam Set : A set of QoS param et ers for which t he BS and possibly t he SS are reserving resources since t he associat ed service flows have been adm it t ed by t he BS. Resources are not lim it ed t o bandwidt h and can include resources such as m em ory. Act iveQoSParam Set : A set of QoS param et ers t hat reflect t he act ual service being provided t o t he associat ed act ive service flows.

Figu r e 7 .1 0 . QoS Pr ovision M ode l

The st andard assum es an Aut horizat ion Module in t he BS t hat approves or denies every change of QoS param et ers associat ed wit h a service flow. This m odule can provision a service flow im m ediat ely or can defer t he service flow act ivat ion t o a lat er period. The st andard present s an " envelope" m odel t hat lim it s t he possible values of t he Adm it t edQoSParam Set and Act iveQoSParam Set . The st andard recognizes t wo m odels: Provisioned Aut horizat ion Model and Dynam ic Aut horizat ion Model. I n t he Provisioned Aut horizat ion Model t he param et ers are provided, for exam ple, by t he net work m anagem ent syst em . I n t he Dynam ic Aut horizat ion Model, t he aut horizat ion m odule issues it s decisions based on it s vendor specific im plem ent at ion. This im plem ent at ion m ay use rout ines t hat com m unicat e t o ext ernal policy servers for addit ional inform at ion and input . The relat ionships

am ong t he QoS Param et er Set s in each m odel are shown in Figure 7.10. The Act iveQoSParam Set is always a subset of t he Adm it t edQoSParam Set , which is always a subset of t he aut horized " envelope." I n t he Dynam ic Aut horizat ion Model, t his envelope is det erm ined by t he Aut horizat ion Module ( labeled as t he Aut horizedQoSParam Set ) . I n t he Provisioned Aut horizat ion Model, t his envelope is det erm ined by t he ProvisionedQoSParam Set .

7.5.2 QoS Mechanisms I n t his sect ion, we revisit t he service flow and it s m echanism s, em phasizing t he QoS aspect s.

7.5.2.1 Classification All packet s generat ed by act ive applicat ions are t agged wit h CI D and SFI D. The classificat ion m odule ident ifies t he packet s based on t hese t ags and forwards t hem t o t he corresponding queues ( see Figure 7.11) .

Figu r e 7 .1 1 . Cla ssifica t ion M odu le

I ndividual applicat ions can est ablish individual connect ions. I EEE 802.16 can provide per- flow classificat ion which support s per- flow QoS services.

7.5.2.2 Channel Access As described in Sect ion 7.5.1, t he channel access m et hod uses TDM for t he downlink and TDMA for t he uplink. The channel access is cont rolled by t he BS t hrough t he UL- MAP and DL- MAP. Bot h TDM and TDMA are collision- free channel access schem es which provide t ight channel access cont rol. Therefore, t hese schem es can provide QoS support for applicat ions wit h st rict QoS requirem ent s.

7.5.2.3 Packet Scheduling The packet scheduling m odule allocat es bandwidt h for connect ions in t erm s of t he num ber of t im e slot s allocat ed per connect ion on t he TDM channel. This m odule also det erm ines when a connect ion is allowed t o t ransm it t he dat a. This inform at ion is defined in t he UL- MAP and DLMAP. There are t wo packet scheduling m odules: uplink packet scheduling and downlink packet scheduling. The downlink packet scheduling m odule is relat ively sim ple because all t he queues reside in t he BS. Therefore, t he downlink packet scheduling m odule can easily ret rieve t he st at e of t he queues. The st andard does not define t he downlink packet scheduling algorit hm . The downlink packet scheduling algorit hm s can be sim ilar t o t he scheduling algorit hm s used in a rout er. The uplink packet scheduling m odule is m ore com plex because t he queues are dist ribut ed am ong m ult iple SSs. The uplink packet scheduling m odule obt ains t he st at e of t he queues and t he bandwidt h requirem ent s from each connect ion t hrough t he BW Request m essage as defined in t he st andard. The uplink packet scheduling det erm ined by t he uplink packet scheduling algorit hm is reflect ed in t he UL- MAP. Figure 7.12 illust rat es t he uplink scheduling process.

Figu r e 7 .1 2 . Uplin k Pa ck e t Sch e du lin g Pr oce ss

The st andard does not define t he scheduling algorit hm t hat det erm ines t he UL- MAP. The uplink scheduling algorit hm has t wo m odes of allocat ing bandwidt h, GPC and GPSS: ●

●

I n GPC m ode, t he uplink scheduling algorit hm dist ribut es t he bandwidt h on a perconnect ion basis. The int elligence of t he uplink scheduling algorit hm resides in t he BS. The packet scheduling m odule at each SS is sim ple because it j ust follows t he UL- MAP. I n GPSS, t he uplink packet scheduling dist ribut es bandwidt h on a per- st at ion basis. The SS dist ribut es it s allocat ed bandwidt h t o each connect ion wit hin t he SS. Com pared t o GPC, t he GPSS based uplink packet scheduling algorit hm t hat resides at t he BS is less com plex. On t he ot her hand, t he packet scheduling algorit hm t hat resides at each SS is m ore com plex.

Figure 7.13 sum m arizes I EEE 802.16 QoS archit ect ure which can deliver m ult iple levels of QoS service: Quant it at ive Service, Qualit at ive Service, and BE Service.

Figu r e 7 .1 3 . I EEE 8 0 2 .1 6 QoS Ar ch it e ct u r e

< Day Day Up >

< Day Day Up >

7.6 IEEE 802.16a The I EEE 802.16a group is considering an ext ension t o t he st andard. This sect ion provides a glim pse int o t he t hought s and ideas discussed in t he group. The final ext ension, when approved, m ay be significant ly different t han t hese t hought s. I EEE 802.16a focuses on ext ending t he st andard support t o a m esh net work t opology in which direct line of sight is not required and in which SSs can com m unicat e wit h each ot her ( see Figure 7.14) . This m esh t opology is proposed in addit ion t o t he point t o m ult i- point t opology covered in t he approved st andard. Anot her addit ion is ext ending t he physical layer t o cover frequencies bet ween 2 and 11 GHz in addit ion t o t he 10 t o 66 GHz covered in t he approved st andard. The MAC includes t he addit ion of an ARQ ( Aut om at ic Repeat Request ) m echanism for handling com m unicat ion errors as well as furt her signaling support for t raffic scheduling algorit hm s t hat provide QoS support .

Figu r e 7 .1 4 . I EEE 8 0 2 .1 6 a M e sh N e t w or k Topology

A m esh net work t opology is a m uch m ore com plex archit ect ure t han t he point t o m ult ipoint t opology which ext ends t he geographical coverage of t he net work. I n a m esh t opology, each SS can com m unicat e wit h anot her SS, and each SS can forward t raffic t ransm it t ed bet ween t he BS and t he dest inat ion SS. The support of a m esh t opology is m ore com plex since we need t o consider issues such as rout ing, collision avoidance and resolut ion bet ween adj acent SSs, synchronizat ion, and m ore. The 2 t o 11 GHz band considered by I EEE 802.16 provides a physical environm ent in which line of sight is not required. This is due t o t he longer wavelengt h as com pared wit h t he 10 t o 66 GHz physical environm ent specified in t he approved st andard. The channel bandwidt hs used in

t his proposed ext ension are bet ween 1.5 and 28 MHz. This addit ion proposes t hree radio t echnologies: single- carrier m odulat ion form at , OFDM ( Ort hogonal Frequency Division Mult iplexing) , and OFDMA ( Ort hogonal Frequency Division Mult iple Access) . OFDMA is a version of OFDM in which t ransm issions are carried in sim ult aneous subchannels com pared t o one channel in OFDM. OFDMA allows for increased bandwidt h as well as increased resiliency t o errors. The proposal det ails com m unicat ion in t he 2 t o 11 GHz frequencies where t here are t wo licensefree bands, 2.4 GHz and 5.4 GHz, t hat were allocat ed prim arily for WLAN users. To avoid int erfering wit h such users, referred t o as Prim ary Users, t he proposal defines algorit hm s t o det ect and avoid int erfering wit h WLAN users. Referred t o as Dynam ic Frequency Select ion ( DFS) , t his process is an abilit y of a syst em t o swit ch t o different physical RF channels bet ween t ransm it and receive act ivit y based on channel m easurem ent crit eria. The DFS procedures t est channels for prim ary users discont inue com m unicat ion once such prim ary users are det ect ed, and periodically schedule for channel t est ing t o det ect new prim ary users.

7.6.1 IEEE 802.16a MAC Discussions The proposal for t he m esh t opology suggest s bot h dist ribut ed and cent ralized scheduling m echanism s. Each m echanism can be em ployed wit h direct ional ant ennas, adapt ive ant enna syst em s, or regular om ni- direct ional ant ennas. I n t he m esh t opology, SSs t hat j oin t he net work do not need t o int eract direct ly wit h t he BS but rat her wit h t heir neighbor SSs. I n dist ribut ed scheduling, all t he SSs need t o coordinat e t heir t ransm issions wit h t heir neighbors t hat are up t o t wo hops away. They broadcast t heir available resources, request s, and grant s t o all of t heir neighbors. They can also schedule t ransm issions by using direct ed request s and grant s. The proposal does not different iat e in t he scheduling m echanism bet ween downst ream and upst ream channels. However, SSs need t o ensure t hat t heir result ing t ransm issions do not collide wit h already scheduled dat a and cont rol t raffic. Dist ribut ed scheduling can be execut ed in eit her a coordinat ed or an uncoordinat ed m anner. The coordinat ed schem e em ploys scheduling packet s t hat are t ransm it t ed in collision- free, regular periods wit hin scheduling cont rol subfram es. The uncoordinat ed schem e allows cont ent ion based access while avoiding conflict s wit h t he schedules est ablished using t he coordinat ed schem e. I n case of collisions, t he SS backs off and t ries t o ret ransm it . I n cent ralized scheduling t he BS allocat es all resources. The BS collect s t he resource request s from all t he SSs t hat are up t o a predet erm ined hop count from t his BS. Using t his inform at ion, t he BS grant s resources for each com m unicat ion link for bot h downst ream and upst ream channels. The BS announces it s decision t o all of t he SSs wit hin it s hop range. SSs t hat receive eit her request s from ot her SSs or grant s from t he BS forward t hese m essages t o ot her neighbor SSs such t hat all part icipant s wit hin t he m esh are covered. To avoid m ult iple redundant t ransm issions of t he sam e dat a, t he m echanism assum es t he exist ence of an algorit hm where SSs know t heir hop count from t he BS. The proposal suggest s t hat t he grant m essages will not cont ain t he act ual grant ed schedule but rat her param et ers from which each SS can com put e it s allocat ed resources using a predet erm ined algorit hm . < Day Day Up >

< Day Day Up >

Part 4: Wireless Personal Area Networks

< Day Day Up >

< Day Day Up >

Chapter 8. Bluetooth Sect ion 8.1. I nt roduct ion Sect ion 8.2. Archit ect ure Sect ion 8.3. Physical Layer Sect ion 8.4. Bluet oot h Baseband Sect ion 8.5. Link Manager ( LM) Sect ion 8.6. Host Cont rol I nt erface ( HCI ) Sect ion 8.7. Logical Link Cont rol and Adapt at ion Prot ocol ( L2CAP) Sect ion 8.8. Higher Bluet oot h Layers Sect ion 8.9. Profiles Sect ion 8.10. QoS Support < Day Day Up >

< Day Day Up >

8.1 Introduction The Bluet oot h st andard for wireless personal area net works is defined by t he Bluet oot h Special I nt erest Group ( SI G) originally founded by Ericsson, I BM, I nt el, Nokia, and Toshiba. Present ly, Bluet oot h SI G also includes com panies such as 3Com , Agere, Microsoft , Mot orola, and hundreds of ot her m em ber com panies. Bluet oot h is nam ed aft er Denm ark's first king Harald Blaat and ( Bluet oot h in English) ( 940- 981) , son of Gorm t he Old King of Denm ark and Thyra Danebod. Som e suggest t hat t he king liked eat ing blueberries and t hus his t eet h becam e st ained wit h t he blue color. Harald Blaat and unit ed Denm ark and Norway. This inspired t he nam e " Bluet oot h," m eaning " unit ing" devices t hrough Bluet oot h. Bluet oot h was first int roduced by Ericsson as a pract ical t echnology for sm all form fact or, inexpensive, short - range radio links bet ween PCs, handheld devices, m obile phones, and ot her com put ing and elect ronic devices t hat com m unicat e and int erface wit h cellular phones and t he I nt ernet . Many com panies have already int roduced Bluet oot h com pliant product s. First we would like t o com pare Bluet oot h wit h infrared t echnology. The I nfrared Dat a Associat ion ( I rDA) has int roduced t wo st andards: I rDA- Dat a, for high- speed, short - range, lineof- sight , and point - t o- point dat a t ransfer, and I rDA- Cont rol, for lower speed com m unicat ion such as wireless keyboards and j oyst icks. The range of I rDA is around 1 m et er and requires line of sight bet ween t he t ransm it t er and receiver. Moreover, I rDA com m unicat ion is subj ect ed t o light int erferences. The biggest advant age of I rDA over Bluet oot h for point - t o- point com m unicat ion is it s high t hroughput , which m akes it suit able for m ult im edia applicat ions requiring high- speed t ransm issions. However, Bluet oot h has larger range, does not require line of sight , and is not prone t o light int erferences. Bluet oot h 1.1 specifies a wireless t echnology in t he 2.4 GHz I ndust rial, Scient ific, and Medical ( I SM) band t hat support s a range of 10 m et ers. Operat ing on t he unlicensed I SM band, Bluet oot h requires t he use of spread spect rum t ransm ission t echnology which is resist ant t o int erference. Bluet oot h deploys fast rat e frequency hopping ( com pared t o ot her Frequency Hopping Spread Spect rum ( FHSS) wireless t echnologies) which provides robust dat a t ransm ission. I t support s sim ult aneous t ransm issions of bot h voice and dat a. Up t o eight dat a devices can be connect ed in a cell, referred t o as piconet . Up t o 10 cells or piconet s can exist wit hin a 10 m et er range. These cells are arranged in a m ult i- cell, referred t o as scat t ernet , which increases Bluet oot h's coverage range t o around 100 m et ers. Each piconet can support up t o t hree sim ult aneous full duplex voice devices. Bluet oot h 1.1 st andard uses a com binat ion of circuit and packet swit ching. I t can support an asynchronous dat a channel, up t o t hree sim ult aneous synchronous voice channels, or a channel which sim ult aneously support s asynchronous dat a and synchronous voice. Each voice channel support s a 64 kbps synchronous voice channel in each direct ion. The asynchronous channel can support four asym m et ric channels wit h dat a rat es of up t o 723 kbps and 57 kbps in t he ret urn direct ion, or 434 kbps for a sym m et ric channel. The earlier Bluet oot h 1.0 version had problem s relat ed t o int eroperabilit y. For exam ple, Bluet oot h uses an FHSS radio t echnology in 2.4 GHz frequency band wit h 79- hop frequency hopping sequences. To m aint ain com m unicat ion, t he st at ions have t o synchronize on t he sam e hopping sequence. Som e European count ries allowed only 23- hop frequency hopping sequences, which creat ed an int eroperabilit y issue. The newer Bluet oot h 1.1 specificat ion addressed t his issue as well as m any ot her int eroperabilit y issues. Fut ure Bluet oot h 2.0 st andard discusses higher speeds, im proved funct ionalit y, different radios, support for ad hoc peer- t o- peer net works, support for devices not current ly included such as

keyboards, support for high- end m ult im edia such as video and m usic, and m ore. < Day Day Up >

< Day Day Up >

8.2 Architecture 8.2.1 Network Topology Bluet oot h defines t hree t opologies: point - t o- point ( Figure 8.1) , single cell ( piconet ) ( Figure 8.2) and m ult i- cell ( scat t ernet ) ( Figure 8.3) . I n a point - t o- point archit ect ure t wo devices com m unicat e direct ly while one device becom es t he m ast er and t he ot her becom es t he slave. I n a piconet , up t o eight act ive devices com m unicat e while one device becom es t he m ast er and t he rest becom e slaves. Many m ore devices can be in " parked" m ode, m eaning non- act ive and occasionally list ening t o t he m ast er for synchronizat ion and broadcast m essages. The m ast er defines t he frequency hopping sequence and t he clock for all t he devices in t he piconet . A scat t ernet can connect up t o 10 piconet s. The connect ion bet ween t he piconet s is handled by a device ( eit her m ast er or slave) t hat part icipat es in bot h piconet s. For exam ple, t he device can serve as a slave in bot h piconet s or t he device can serve as t he m ast er in one piconet as well as a slave in anot her piconet . To avoid collisions and int erference bet ween piconet s, each piconet uses a different frequency hopping series which is uniquely defined by t he m ast er. However, t he m ore piconet s are act ive in t he sam e area, t he higher t he chance t hat such piconet s will t ransm it on t he sam e frequency, leading t o m ore collisions and int erference.

Figu r e 8 .1 . Poin t - t o- Poin t Topology ( sin gle sla ve ope r a t ion )

Figu r e 8 .2 . Picon e t

Figu r e 8 .3 . Sca t t e r n e t

8.2.2 Protocol Stack As shown in Figure 8.4, t he st andard det ails various prot ocol layers: t he Bluet oot h Radio, t he Bluet oot h Baseband, t he Link Manager ( LM) , t he Logical Link Cont rol and Adapt at ion Prot ocol ( L2CAP) , t he Host Cont rol I nt erface ( HCI ) , t he Service Discovery Prot ocol ( SDP) , t he Audio/ Telephony Cont rol prot ocol Specificat ion ( TCS) , t he Radio Frequency orient ed em ulat ion of t he serial com m unicat ion port s ( RFCOMM) , t he Hum an I nt erface Device ( HI D) , TCP/ I P, and ot her high- level prot ocols.

Figu r e 8 .4 . Blu e t oot h Pr ot ocol St a ck

The Bluet oot h Radio layer, which is parallel t o t he physical layer in t he seven- layer Open Syst em s I nt erconnect ion ( OSI ) m odel, defines t he requirem ent s for a 2.4 GHz Bluet oot h t ransceiver ( i.e., frequency band, radio param et ers, t ransm it t er and receiver charact erist ics) . The Baseband layer describes t he specificat ions of t he Bluet oot h Link Cont roller which execut es t he baseband prot ocols and ot her low- level link funct ions. The Baseband layer includes t he following m odules: LM, HCI , and L2CAP. The LM prot ocol specifies t he link set up and cont rol. The HCI provides a com m and int erface t o t he Baseband Link Cont roller and LM and access t o t he hardware st at us and cont rol regist ers. The L2CAP support s higher level prot ocol m ult iplexing, packet segm ent at ion, and packet reassem bly, and it conveys QoS inform at ion. The RFCOMM prot ocol provides em ulat ion of serial port s over t he L2CAP prot ocol. The Service Discovery Prot ocol ( SDP) provides ways for applicat ions t o discover which services are available. < Day Day Up >

< Day Day Up >

8.3 Physical Layer The Bluet oot h radio uses FHSS in 79 frequencies, each in a 1 MHz range from 2.402 GHz t o 2.480 GHz. The radio t ransceiver changes t he hopping frequency in a pseudorandom fashion det erm ined by t he m ast er. The hopping rat e is 1600 hops per second. GFSK ( Gaussian Frequency Shift Keying) is used for t he m odulat ion schem e. Each device is classified int o power classes 1, 2, and 3. Power class 1 is for long- range t ransm ission of around 100 m et ers. Power class 2 is for short - range t ransm ission of around 10 m et ers wit h less power out put t han power class 1. Power class 3 is for short - range t ransm ission of around 10 cm , using m inim al t ransm ission power. The power cont rol funct ion regulat es t he t ransm ission power for opt im izing t he power consum pt ion and channel int erference level. < Day Day Up >

< Day Day Up >

8.4 Bluetooth Baseband The Baseband execut es t he key funct ions of t he Media Access Cont rol ( MAC) layer and Dat a Link Cont rol ( DLC) layer. The MAC funct ions include channel access and packet handling t hrough asynchronous and synchronous links. The link cont rol prot ocol of t he Baseband is im plem ent ed as a Link Cont roller, which works wit h t he LM for carrying out link level rout ines such as link connect ion and power cont rol.

8.4.1 Channel Definition and Channel Access A piconet consist s of a m ast er node, act ive slave nodes ( up t o seven nodes) , and several inact ive slave nodes ( up t o 200 nodes) . All connect ions ( or links) in t he piconet t ake place only bet ween t he m ast er and slaves in point - t o- point or point - t o- m ult ipoint fashion. There is no direct connect ion bet ween slaves. The com m unicat ion channel is st ruct ured int o t im e slot s, each wit h durat ion of 625 µs. The m ast er uniquely defines t he frequency hopping sequence based on it s device address ( 48 bit s) . Generally, all devices hop or change t he frequency every t im e slot ( wit h m ax hopping rat e of 1600 hops per second, which corresponds t o 625 µs t im e slot durat ion) except in t he case of m ult i- slot packet t ransm ission, which we will describe lat er in t his sect ion. All slaves ( bot h act ive and inact ive) in a piconet are required t o synchronize t he clock and frequency hopping sequence wit h t he m ast er. A packet t ransm ission st art s at t he beginning of a t im e slot and can last for one, t hree, or five t im e slot s. Bluet oot h deploys a full duplex, Tim e Division Duplex ( TDD) prot ocol in which t ransm ission and recept ion alt ernat e. By alt ernat ing t im e slot s, t he m ast er and slave can t ransm it packet s according t o t he TDD prot ocol. The m ast er st art s it s t ransm ission only in even num bered t im e slot s and t he slave st art s it s t ransm ission only in odd num bered t im e slot s. Figure 8.5 illust rat es a single t im e slot packet t ransm ission.

Figu r e 8 .5 . Ex a m ple of a Sin gle Tim e Slot Tr a n sm ission Usin g TD D

A single- slot packet is t ransm it t ed wit h a hopping frequency of t he current Bluet oot h clock value. I n case of a m ult i- slot packet , t he packet will be t ransm it t ed wit h t he hopping frequency

of t he first t im e slot for t he ent ire durat ion of packet t ransm ission. The m ult iple t im e slot t ransm ission is shown in Figure 8.6.

Figu r e 8 .6 . M u lt i- Slot Pa ck e t s

Bluet oot h aim s t o support bot h voice and dat a t ransm ission by using a com binat ion of bot h circuit and packet swit ching prot ocols. Bluet oot h Baseband defines t wo t ypes of t ransm ission links bet ween a m ast er and a slave: synchronous connect ion- orient ed ( SCO) link and asynchronous connect ionless ( ACL) link. SCO is used for synchronous packet t ransm ission while ACL is used for asynchronous dat a t ransm ission.

8.4.1.1 SCO and ACL links The SCO link is a sym m et ric, point - t o- point link bet ween t he m ast er and a single slave in t he piconet . The SCO link is est ablished by t he m ast er by sending an SCO set up packet via t he LM prot ocol. This packet cont ains t im ing param et ers such as reserved t im e slot s. The m ast er m aint ains t he SCO link by reserving t im e slot s at regular int ervals—t hat is, est ablishing a circuit swit ched connect ion. A slave can support up t o t hree SCO links from t he sam e m ast er or t wo SCO links if t he links originat e from different m ast ers. The m ast er can support up t o t hree sim ult aneous SCO links t o t he sam e slave or t o different slaves. Wit h t he reserved t im e slot s at regular int ervals, t he SCO link provides bounded delay channel access which is suit able for t im e sensit ive applicat ions such as voice. The m ast er sends packet s, referred as SCO packet s, at regular int ervals, called SCO int ervals and count ed in slot s in t he reserved m ast er- t o- slave slot s. The SCO slave is allowed t o respond wit h an SCO packet in t he following reserved slavet o- m ast er slot . Figure 8.7 shows t wo SCO links wit h different SCO int ervals. Bandwidt h allocat ion t o SCO links and channel access delay are det erm ined by t he SCO int erval.

Figu r e 8 .7 . SCO Lin k s

There is no packet ret ransm ission in SCO t ransm issions. The error handling relies on t he forward error correct ing ( FEC) schem e and t he codec used. SCO is designed t o be used by voice packet t ransm issions. Bluet oot h specificat ion cat egorizes t he voice packet int o t hree t ypes: HV1, HV2, and HV3. HV st ands for high- qualit y voice. The differences bet ween t he t hree t ypes of HV are t he num ber of inform at ion byt es and t he error correct ing schem e. HV1 provides t he lowest num ber of inform at ion byt es ( low dat a rat e) but uses t he m ost robust error correct ion schem e ( i.e., 1/ 3 FEC) . HV3 provides t he highest num ber of inform at ion byt es ( high dat a rat e) but does not use any error correct ion schem e. While achieving high dat a rat e, HV3 is suscept ible t o channel int erference. Error handling for HV3 only relies on t he codec of voice cont ent . The ACL link is a point - t o- m ult ipoint link bet ween t he m ast er and all t he slaves in t he piconet . I n t he slot s not reserved for SCO packet s, t he m ast er can est ablish an ACL link on a per- slot basis ( i.e., packet swit ched connect ion) t o any slaves, including t he slave( s) already engaged in an SCO link. Only a single ACL link bet ween a m ast er and a slave can exist . That m eans t hat m ult iple packet swit ched connect ions on a slave are aggregat ed t o t he sam e single ACL link. ACL support s bot h asynchronous and isochronous services. For m ost ACL packet s, packet ret ransm ission is support ed t o ensure int egrit y. A slave can respond t o an ACL packet in t he slave- t o- m ast er slot only if it has been addressed in t he preceding m ast er- t o- slave slot . ACL packet s t hat are not addressed t o a specific slave are considered as broadcast packet s. The slave's t ransm ission of an ACL packet is cont rolled by a POLL packet sent from t he m ast er in t he m ast er- t o- slave slot . When receiving t he POLL packet , t he slave responds by sending dat a packet s in case it has packet s t o send, or in case it has not hing t o send, it sends a NULL packet in t he following slave- t o- m ast er slot . The bandwidt h allocat ion for a slave's ACL link is cont rolled by t he polling frequency of t he m ast er. The m ore frequent ly a slave receives poll packet s, t he m ore bandwidt h a slave is allocat ed. Bluet oot h specificat ions define various t ypes of asynchronous dat a packet s for t he ACL link ( i.e., DM1, DH1, DM3, DH3, DM5, DH5, AUX) which have different payload size, error det ect ion schem e, error correct ing schem e, and achievable dat a rat e. The ACL link is used by asynchronous dat a packet s and cont rol packet s ( i. e., from LC, LM, L2ACP) .

8.4.1.2 Logical Channels Bluet oot h defines five logical channels used t o t ransfer different t ypes of inform at ion: ●

●

●

●

●

Link Cont rol ( LC) Channel: This channel carries low- level link cont rol inform at ion such as flow cont rol and payload charact erizat ion. Link Manager ( LM) Channel: Typically it carries cont rol inform at ion exchanged bet ween t he m ast er's LMs and t he slaves. User Asynchronous ( UA) Channel: I t carries L2CAP t ransparent asynchronous user dat a. These dat a m ay be t ransm it t ed in one or m ore baseband packet s. User I sochronous ( UI ) Channel: I t carries L2CAP t ransparent user isochronous dat a. These dat a m ay be t ransm it t ed in one or m ore baseband packet s. User Synchronous ( US) Channel: I t carries t ransparent synchronous user dat a. This channel is carried over t he SCO link.

LC and LM are used at t he link cont rol level and LM. UA, UI , and US are used t o carry user inform at ion. The inform at ion on t he LC channel is carried in t he packet header, while t he inform at ion for all ot her channels is carried in t he packet payload. The inform at ion for t he US channel is carried only by t he SCO link. The inform at ion for t he UA and UI channels is norm ally carried by t he ACL link; however, it can also be carried by t he dat a port ion of t he com bined dat a- voice packet on t he SCO link. The inform at ion on t he LM channel can be carried eit her by t he SCO or t he ACL link. Figure 8.8 shows t he flow of each logical channel.

Figu r e 8 .8 . Logica l Ch a n n e ls

8.4.1.3 Packet Format Figure 8.9 shows t he st andard packet form at , which includes t he Access Code, LC header, and payload.

Figu r e 8 .9 . St a n da r d Pa ck e t For m a t

There are t hree t ypes of access codes: ●

Channel Access Code ( CAC) : CAC is used t o ident ify a piconet . CAC is det erm ined by t he device address of t he piconet 's m ast er. Therefore, CAC is unique in each piconet . CAC is norm ally included in t he user dat a packet .

●

●

Device Access Code ( DAC) : DAC is used t o ident ify a device. DAC is uniquely det erm ined by t he device address. DAC is used for t he paging procedures. For exam ple, in t he paging procedure, t he m ast er will send t he DAC of t he slave wit h which it want s t o est ablish a connect ion. Then, t he slave responds back wit h it s DAC. The DAC is included in t he paging and paging response m essages which include only t he access code field ( no header and payload) . I nquiry Access Code ( I AC) : I AC is used for t he inquiry procedures. I AC is included in t he inquiry m essage which includes only t he access code field ( no header and payload) . The m ast er sends inquiry m essages t o collect t he device addresses of slaves in t he t ransm ission range.

Figure 8.10 shows t he packet form at of t he ACL and SCO links. An SCO packet cont ains t he voice payload while an ACL packet cont ains t he dat a payload. Som e SCO packet s can also cont ain dat a payload. The LC can use eit her SCO or ACL packet wit h or wit hout payload ( i.e., NULL and POLL packet s do not have dat a payload while frequency hop synchronizat ion ( FHS) packet s include cont rol inform at ion in t he dat a payload) . The LM includes it s cont rol inform at ion in t he dat a payload.

Figu r e 8 .1 0 . SCO a n d ACL Pa ck e t For m a t

8.4.2 Link Control Before t he m ast er and slaves can com m unicat e wit h each ot her, t hey have t o est ablish t he piconet using link cont rol m echanism s based on link cont rol st at es. Link cont rol inform at ion is included in t he link cont rol header of t he packet as shown in Figure 8.10.

8.4.2.1 Link Control States A st at ion can be in one of a num ber of link cont rol st at es ( see Figure 8.11) . There are t wo m aj or st at es: St andby and Connect ion. I n addit ion, t here are subst at es wit h t heir associat ed procedures: page, page scan, inquiry, and inquiry scan. These procedures are used t o est ablish a piconet or add new slaves t o a piconet . The subst at es are t ransient st at es bet ween t he St andby and Connect ion st at es.

Figu r e 8 .1 1 . Lin k Con t r ol St a t e D ia gr a m

8.4.2.1.1 Standby State The st at ion default st at e is St andby st at e, which is a low- power consum pt ion m ode. The St andby st at ions are not associat ed wit h any piconet s. Occasionally, t he Link Cont roller m ay leave t his st at e t o scan for page or inquiry m essages, or t o page or inquiry it self. I f t he st at ion received such a m essage ( i.e., page m essage, inquiry m essage) , it will ent er t he Connect ion st at e and becom e a slave. I f t he st at ion it self t ransm it s a successful page m essage or inquiry m essage, t hen it will ent er t he Connect ion st at e and becom e t he m ast er. I nquiry procedures ( see Figure 8.12) are t he int eract ion processes bet ween t he st at ion in t he inquiry subst at e ( called inquiry st at ion) and t he st at ion in t he inquiry scan subst at e ( inquiryscan st at ion) . I nquiry is used where t he dest inat ion's device address is unknown t o t he source. The inquiry st at ions are t he st at ions t hat want t o discover which new st at ions are wit hin t heir t ransm ission range. The inquiry- scan st at ions are t he st at ions t hat want t o be discovered by ot her st at ions. The inquiry st at ion t ransm it s t he inquiry m essage ( cont aining I AC) wit h a frequency t hat follows t he inquiry frequency hopping sequence. The inquiry- scan st at ion scans t he inquiry m essage at a single hop frequency. The inquiry- scan st at ions can receive t he inquiry m essage whenever t he inquiry st at ion t ransm it s at t he sam e frequency. Aft er receiving t he inquiry m essage successfully, t he inquiry- scan st at ion will send t o t he inquiry st at ion t he response m essage t hat cont ains t he device address and t he clock. I n sum m ary, during t his process, t he inquiry st at ion collect s t he device addresses and clocks of all st at ions t hat respond t o t he inquiry m essage. Based on t his inform at ion t he inquiry st at ion can issue a page m essage t o est ablish connect ion t o known st at ions.

Figu r e 8 .1 2 . I n qu ir y Se qu e n ce s

Paging procedures are used t o est ablish new connect ions bet ween t he m ast er and a slave. The paging subst at e is deployed by t he m ast er t o est ablish t he new connect ion wit h t he slave which is in t he paging- scan subst at e. The paging procedures are sim ilar t o t he inquiry procedures. The m ast er t ries t o connect t o t he t arget ed slave by repeat edly t ransm it t ing t he slave's device access code ( DAC) in several hop channels. Since t he clocks of t he m ast er and t he slave are not yet synchronized, t he m ast er does not know exact ly when t he slave wakes up and on which hop channel. The m ast er predict s t he hop channels where t he slave should be, for exam ple, by recalling det ails of t he last exchange of packet s. Based on t his predict ion, it t ransm it s a series of t he sam e DACs at t he predict ed channel as well as ot her channels and list ens bet ween t ransm issions for a response from t he slave. I n Page Scan, a slave list ens for it s own DAC for t he durat ion of t he scan window. During t his window, t he slave list ens at a single hop frequency, it s correlat or m at ched t o it s device access code. When a slave ent ers Page Scan, it select s t he scan frequency according t o t he page hopping sequence corresponding t o t his st at ion's device address. Figure 8.13 depict s t he paging sequences bet ween t he m ast er and a slave. The m ast er t ransm it s paging m essages m ore frequent ly ( 3200 packet s per second or once every 312.5 µs) t han t he norm al dat a packet t ransm issions ( 1600 packet s per second or once every 625 µs) . The m ast er det erm ines it s paging frequency hopping sequence based on it s device address. The slaves list en or scan at a fixed frequency. I f a slave receives t he paging m essage, it will t ransm it a paging response m essage which cont ains it s DAC. Then, t he m ast er t ransm it s t he FHS m essage in order for t he slave t o synchronize 1) t o t he channel frequency hopping sequence and 2) t o t he clock. Then, t he slave replies wit h a FHS packet and becom es an act ive node of t he piconet .

Figu r e 8 .1 3 . Pa gin g Se qu e n ce s

I n St andby, no connect ion has been est ablished and t he st at ion can use all of it s capacit y t o support Page Scan. I f desired, t he st at ion can place ACL connect ions in t he Hold m ode or even use t he Park m ode before ent ering Page Scan. SCO connect ions are preferably not int errupt ed by Page Scan. I n t his case, Page Scan m ay be int errupt ed by t he reserved SCO slot s which have higher priorit y t han Page Scan.

8.4.2.1.2 Connection State At t he connect ion st at e, a Bluet oot h st at ion becom es t he m em ber of a piconet . The st at ion can be in one of t he following four connect ion m odes: Act ive, Sniff, Hold, and Park. These connect ion m odes are used t o save power and t o allow st at ions t o com m unicat e wit h different piconet s. Act ive Mode: I n t his m ode t he st at ion can act ively com m unicat e on t he wireless channel. A piconet is allowed t o have up t o eight act ive m ode st at ions at t he sam e t im e. The m ast er schedules t he t ransm ission based on t raffic requirem ent s t o and from t he slaves. I n addit ion, t he m ast er support s regular t ransm issions t o keep slaves synchronized t o t he channel. Act ive slaves list en during t he m ast er- t o- slave slot s. I f an act ive slave is not addressed, it m ay sleep unt il t he next new m ast er t ransm ission. A periodic m ast er t ransm ission is required t o keep t he slaves synchronized t o t he channel. Since t he slaves only need t he channel access code for synchronizat ion, any packet t ype can be used for t his purpose. Sniff Mode: Slave st at ions can change t o a power- saving m ode in which t he st at ion's act ivit y is decreased. I n t his Sniff m ode t he slave st at ion list ens t o t he piconet less oft en, t hus reducing power consum pt ion. I f a slave part icipat es on an ACL link, it has t o list en t o t he m ast er t raffic every ACL slot . The t im e int erval bet ween consecut ive list ening event s is program m able by t he applicat ion. Hold Mode: Slave st at ions can change int o anot her power- saving m ode. During t he Connect ion st at e, t he ACL link t o a slave can be in Hold m ode. This m eans t hat t he slave t em porarily does not support ACL packet s while cont inuing t o support possible SCO links. While in Hold m ode, t he st at ion can execut e procedures such as scan, page, inquire, or com m unicat e wit h anot her piconet . Before t he slave ent ers Hold m ode, t he m ast er and slave have t o agree on t he t im e durat ion t hat t he slave can rem ain in t his m ode. Once t his t im e expires, t he st at ion will ret urn t o Act ive m ode, synchronize t o t he t raffic, and wait for inst ruct ions from t he m ast er.

Park Mode: I n t his m ode t he slave st at ion is st ill synchronized t o t he piconet but does not com m unicat e. Parked st at ions give up t heir act ive m em ber address and only occasionally list en t o t he t raffic from t he m ast er t o synchronize and check for broadcast m essages. To support parked slaves, t he m ast er est ablishes a beacon channel where one or m ore slaves are parked. The m ast er t ransm it s on t his channel at const ant t im e int ervals a series of beacon slot s.

8.4.2.2 Piconet Establishment and Operation The piconet is governed ent irely by it s m ast er. The m ast er's address det erm ines t he frequency hopping sequence and t he channel access code ( t he code t hat uniquely ident ifies t he piconet ) . Since Bluet oot h is based on TDD, every Bluet oot h st at ion has an int ernal syst em clock which det erm ines t he t im ing and hopping of t he radio t ransceiver. This clock has no relat ion t o t he t im e of day and can t herefore be init ialized at any value, does not require adj ust m ent s, and is never t urned off. For synchronizat ion wit h ot her unit s, only offset s are used and added t o t he nat ive int ernal clock. To ensure synchronizat ion, t he m ast er set s t he t im ing of all slaves based on it s clock by t ransm it t ing t o t he slaves it s clock reading. The slaves add an offset value t o t heir nat ive clocks so t hey can be synchronized t o t he m ast er clock. Since t he clocks are freerunning, t he offset s have t o be updat ed regularly. Est ablishing piconet is cont rolled by link cont rol st at e. Figure 8.14 shows t he piconet est ablishm ent sequence.

Figu r e 8 .1 4 . Picon e t Est a blish m e n t Se qu e n ce s

First , t he m ast er execut es t he inquiry process t o obt ain t he device addresses and clocks of t he slaves wit hin t he t ransm ission coverage. Then, t he m ast er perform s t he paging process t o est ablish t he connect ions wit h t he specific slave. Finally, bot h t he m ast er and t he act ive slaves can com m unicat e wit h each ot her.

8.4.2.3 Scatternet Establishment and Operation To increase t he coverage area and t he num ber of support ed st at ions, several piconet s m ay coexist in t he sam e area. Each piconet 's m ast er specifies a different hopping series. I n addit ion, t he packet s carried on t he channels are preceded by different channel access codes as det erm ined by t he m ast er addresses. As t he num ber of coexist ing piconet s increases, t he probabilit y of collisions increases since different piconet s m ay t ransm it on one of t he 79 frequency channels. I f several piconet s coexist in t he sam e area, a st at ion can part icipat e in t wo or m ore piconet s by applying t im e m ult iplexing. I n each piconet , t he st at ion will use t his piconet hopping series, m ast er device address, and proper clock offset . A Bluet oot h st at ion can act as a slave in several piconet s but as a m ast er in only one piconet . A group of connect ed piconet s is referred t o as a scat t ernet . St at ions m ust use t im e m ult iplexing t o swit ch bet ween t he piconet s in which t hey part icipat e. I n case of ACL links, a st at ion can request t o ent er t he Hold or Park m ode so t hat during t his

m ode it can j oin t he ot her piconet . Unit s in t he Sniff m ode m ay have sufficient t im e t o visit anot her piconet bet ween t he sniff slot s. I f SCO links are est ablished, ot her piconet s can only be visit ed in t he non- reserved slot s bet ween t he sniff slot s. Since coexist ing piconet s do not have a m echanism t o synchronize clocks, addit ional guard t im e should be considered for proper swit ching. A m ast er can becom e a slave in anot her piconet by being paged by t he m ast er of t his ot her piconet . On t he ot her hand, a slave part icipat ing in one piconet can be a m ast er in anot her piconet . Since t he paging unit always st art s out as m ast er, a m ast er- slave role exchange will be required. Such an exchange m ay involve a reorganizat ion of piconet s since slaves m ay change from t he current m ast er t o t he new m ast er. This is a com plex process since it requires all st at ions t o resynchronize t heir clocks, t ransit t o t he frequency hopping series of t he new m ast ers, and adj ust t o t he new m ast er device addresses. This reorganizat ion of piconet s should be done wit hout int erfering wit h ongoing com m unicat ion act ivit ies and wit hout involving t he users. The opt ion of set t ing up a com plet ely new set of piconet s will require a long period of t im e. Anot her opt ion is t o have t he new m ast ers ut ilize t heir knowledge of t im ing and hopping sequences t o reduce t he t im e for such an exchange. Because of t he com plexit y of t he piconet reorganizat ion process, it is current ly left out side t he scope of t he Bluet oot h st andard. Therefore, Bluet oot h developers need t o design and im plem ent t heir own exchange algorit hm .

8.4.2.4 Link Supervision Several event s can cause st at ions t o loose t heir com m unicat ion link. Such event s can be power failure, m ovem ent out of range, and int erference from out side sources. Hence, t he st andard defines a process in which t he link is m onit ored by t he m ast er and t he slave using link supervision t im ers for bot h SCO and ACL connect ions. When a st at ion receives a packet wit h a valid slave address, t his t im er is reset . However, when t his t im er reaches a predet erm ined value when t he st at ion is in t he Connect ion st at e, t he connect ion is reset . This t im er t im eout period is negot iat ed at t he Link Managem ent ( LM) such t hat it is longer t han Hold and Sniff periods. I f t he slave is in Park m ode, t hen link supervision is done by un- parking and re- parking t he slave. < Day Day Up >

< Day Day Up >

8.5 Link Manager (LM) The Link Manager is responsible for link set up, securit y, aut hent icat ion, link configurat ion, t im er set up, and ot her cont rol procedures ( Figure 8.15) . The com m unicat ion bet ween t he st at ions' LM m odules is done using t he Link Manager Prot ocol ( LMP) . The LM sends it s cont rol m essage t hrough Dat a Medium rat e ( DM1) or Dat a Voice ( DV) packet s. Due t o t he fact t hat DM1 and DV packet s use reliable links ( using 2/ 3 FEC and CRC) , t here are no explicit acknowledgm ent m essages. LM m essages have higher priorit y t han user dat a.

Figu r e 8 .1 5 . Lin k M a n a ge r

LM support s upper layer applicat ions t hat execut e cont rol algorit hm s such as m anaging t he connect ion st at es ( Park, Hold, Sniff, Act ive) , est ablishing securit y, and support ing QoS. These cont rol algorit hm s m ay be self cont ained or m ay allow input from users. We will discuss m ore about t he QoS aspect s of t he LM prot ocols in a lat er sect ion. < Day Day Up >

< Day Day Up >

8.6 Host Control Interface (HCI) As we showed in Figure 8.4, which describes t he Bluet oot h prot ocol st ack, HCI is t he layer bet ween L2CAP and LM. The HCI provides a st andard int erface t o Bluet oot h so t hat upper layers can be independent and t ransparent from t he host hardware im plem ent at ion. I t provides int erface t o t he LM and access t o t he hardware st at us and cont rol regist ers. HCI consist s of m ult iple part s ( i.e., HCI driver, HCI firm ware) t hat reside in bot h Bluet oot h Host and Bluet oot h hardware. Figure 8.16 shows HCI relat ionship t o t he Bluet oot h syst em .

Figu r e 8 .1 6 . H CI Re la t ion sh ip t o t h e Blu e t oot h Syst e m

< Day Day Up >

< Day Day Up >

8.7 Logical Link Control and Adaptation Protocol (L2CAP) The Logical Link Cont rol and Adapt at ion Layer Prot ocol ( L2CAP) provides connect ion- orient ed and connect ionless dat a services t o upper layer prot ocols. The L2CAP specificat ion is defined for only Asynchronous Connect ionless ( ACL) links. Support for Synchronous Connect ion- Orient ed ( SCO) links, m ainly used for real- t im e voice t raffic wit h reserved bandwidt h, is not yet provided in L2CAP. Bluet oot h recom m ends t hat L2CAP should include sim ple and low- overhead program s so t hey can fit in devices wit h lim it ed com put at ional, power, and m em ory resources. These dat a services include QoS support , group support , and prot ocol m ult iplexing, segm ent at ion and reassem bly. Prot ocol m ult iplexing enables an applicat ion t o sim ult aneously use several higher layer prot ocols such as TCP/ I P and RFCOMM. L2CAP labels each connect ion by a channel ident ifier ( CI D) and assum es t hat each channel is a full duplex connect ion and m ay have a QoS flow specificat ion. The L2CAP connect ion est ablishm ent process allows exchange of t he expect ed QoS bet ween t wo Bluet oot h st at ions. Moreover, each st at ion m onit ors t he resources used t o ensure t hat t he required QoS cont ract is provided. Again, t hese QoS algorit hm s are left t o t he developer and are not provided as part of Bluet oot h. L2CAP group support perm it s higher level applicat ions t o m ap groups ( group of addresses) t o piconet s. The Baseband layer support s t he concept of a piconet for a unique group. Wit hout group support , higher level applicat ions would need t o est ablish a com plex com m unicat ion process direct ly wit h t he Baseband Prot ocol and LM layers. < Day Day Up >

< Day Day Up >

8.8 Higher Bluetooth Layers 8.8.1 RFCOMM RFCOMM enables com pat ibilit y wit h applicat ions t hat use t he serial port as t heir m ain com m unicat ion bus. RFCOMM conveys all of t he RS232 cont rol signals and support s rem ot e port configurat ion.

8.8.2 Service Discovery Protocol (SDP) The service discovery prot ocol ( SDP) provides t ools for applicat ions t o discover which services are available and t o det erm ine t he nat ure of t hese services.

8.8.3 Audio and Telephony Control (Tel Ctrl) This layer includes t he int erface needed t o connect and disconnect a t elephone call, including signaling t he devices t hat part icipat e in t he connect ion. Telephony audio links are est ablished wit h synchronous links and t herefore do not go t hrough t he sam e L2CAP- t o- LM pat h t hat asynchronous links go t hrough. I n ot her words, audio links m ay be t hought of as direct Baseband- t o- Baseband links. < Day Day Up >

< Day Day Up >

8.9 Profiles Bluet oot h profiles specify a set of basic st andards t o ensure device int eroperabilit y. I n Bluet oot h Specificat ions, v. 1.1, t here are 13 Bluet oot h profiles. They include profiles such as Generic Access, Cordless Telephony, I nt ercom , Serial Port , Headset , Dial- up Net working, Fax, and LAN Access and Synchronizat ion ( exchange of inform at ion bet ween applicat ions such as calendar) . < Day Day Up >

< Day Day Up >

8.10 QoS Support Bluet oot h aim s t o support bot h synchronous and asynchronous dat a services. The key concept int roduced by Bluet oot h for QoS support is it s definit ion of t wo dat a link t ypes, SCO and ACL. Furt herm ore, Bluet oot h also defines QoS as enabling MAC and QoS signaling. I n t his sect ion, we t ake a closer look int o Bluet oot h's QoS m echanism s. Figure 8.17 shows Bluet oot h's QoS archit ect ure.

Figu r e 8 .1 7 . Blu e t oot h QoS Ar ch it e ct u r e

8.10.1 Classification Bluet oot h defines packet classificat ion at t he L2CAP layer. All asynchronous dat a packet s are assigned a CI D as shown in Table 8.1.

Ta ble 8 .1 . Ch a n n e l I de n t ifie r s CI D

D e scr ipt ion

0x0000

Null ident ifier

0x0001

Signaling channel

0x0002

Connect ionless recept ion channel

0x0003 – 0x003F

Reserved

0x0040 – 0xFFFF

Dynam ically allocat ed

For a connect ion- orient ed applicat ion, CI Ds are dynam ically assigned on bot h end st at ions. For a connect ionless applicat ion, a fixed CI D ( 0x0002) is assigned t o t he receiver and a dynam ic CI D is allocat ed t o t he sender. For L2CAP signaling, bot h end st at ions are allocat ed CI D= 0x0001. A classifier will use t he CI D t o ident ify t he packet s and forward t hem t o t he appropriat e queue. Bluet oot h does not define how t o classify t he voice packet s t hat do not go t hrough t he L2CAP layer.

8.10.2 Channel Access As described in Sect ion 8.4, Bluet oot h deploys a TDD channel access. The m ast er is t he only st at ion t hat t ransm it s on t he downlink direct ion ( m ast er- t o- slave t im e slot ) . On t he ot her hand, t he uplink ( slave- t o- m ast er t im e slot ) is shared by m ult iple slave st at ions. The uplink channel access is cont rolled by t he m ast er which allocat es reserved t im e slot s for SCO packet s and uses POLL cont rol packet s for ACO packet s. Aft er receiving a POLL, t he slave is allowed t o t ransm it ACO packet s. Bluet oot h uses a collision- free channel access schem e which provides t ight channel access cont rol. Therefore, t hese schem es can provide QoS support for applicat ions wit h st rict QoS requirem ent s.

8.10.3 Packet Scheduling We different iat e bet ween t wo packet scheduling schem es: I nt rast at ion packet scheduling and I nt erst at ion packet scheduling. There are t hree t ypes of packet s: SCO packet s, ACL packet s, and Cont rol packet s ( which use t he ACL link) . I n t he LC m odule of t he slave, a slave can have up t o t hree SCO links for t he sam e m ast er or up t o t wo SCO links for different m ast ers, one ACL link and one cont rol link. I n t he LC m odule of t he m ast er, t he m ast er can have up t o t hree SCO links, one ACL link for each slave, and one cont rol link. Using t hese packet scheduling m echanism s, Bluet oot h can provide st rict QoS service for voice applicat ions using t he SCO link and loose QoS service for asynchronous applicat ions using t he ACL link.

8.10.3.1 Intrastation Packet Scheduling SCO packet s are t ransm it t ed in t he reserved t im e slot and ACL packet s are t ransm it t ed int o t he rest of t he t im e slot s. Therefore, Bluet oot h im plicit ly deploys st rict priorit y packet scheduling. SCO has higher priorit y t han ACL. Furt herm ore, Bluet oot h also explicit ly defines t hat cont rol packet s have higher priorit y t han ACL packet s. I n sum m ary, SCO has t he highest priorit y followed by cont rol packet s and ACL packet s. There is only one ACL link in a slave. Asynchronous dat a packet s from different applicat ions init iat ed by t he sam e slave will be

aggregat ed in t he sam e queue. All packet s experience t he sam e QoS. Moreover, it is possible t hat an asynchronous applicat ion can abuse t he ot her asynchronous applicat ions by using t he bandwidt h as m uch as it can.

8.10.3.2 Interstation Packet Scheduling The slave is allowed t o t ransm it in t he reserved t im e slot ( for SCO) or in response t o a POLL m essage ( for ACL) . The bandwidt h allocat ion for SCO packet s depends on t he num ber of reserved t im e slot s which are assigned by t he signaling process. The bandwidt h allocat ion for ACL packet s depends on t he num ber of POLL packet s received ( polling frequency) which is cont rolled by t he m ast er. Bluet oot h does not define t he polling algorit hm which det erm ines t he polling sequence and polling frequency.

8.10.4 QoS Signaling There are t wo levels of QoS signaling: L2CAP QoS signaling and LM QoS signaling.

8.10.4.1 L2CAP QoS Signaling L2CAP QoS signaling exchanges t he QoS inform at ion bet ween L2CAP of m ast er and slave. Such QoS inform at ion indicat es t he t raffic charact erist ic and QoS requirem ent s of applicat ions ( called FlowSpec) such as Token Rat e, Token Bucket Size, Peak Bandwidt h, and Lat ency and Delay variat ion. The QoS inform at ion is included in t he Configurat ion Param et er Opt ion field of t he L2CAP cont rol m essage. L2CA_ConfigReq is t he L2CAP cont rol m essage t hat perform s t he connect ion request . L2CA_ConfigReq request s t he t raffic charact erist ics and QoS requirem ent of out going dat a flow t o t he m ast er. Then t he m ast er sends L2CA_ConfigRsp in response t o L2CA_ConfigReq. L2CA_ConfigRsp indicat es t he accept ed t raffic charact erist ics of t he incom ing dat a flow. Bluet oot h does not define t he adm ission cont rol process t hat decides t he accept ed t raffic charact erist ics. Bluet oot h does not define t he algorit hm t hat m aps t he FlowSpec t o polling frequency and polling sequence. Anot her L2CAP QoS signaling m essage is L2CA_QoSViolat ionI nd, which is used for indicat ing t he address of t he rem ot e device t hat violat es a QoS cont ract .

8.10.4.2 LM QoS signaling Bluet oot h also defines QoS signaling in t he LM layer. For exam ple, LMP_qualit y_of_service and LMP_qualit y_of_service_req cont ain t he polling int erval for t he ACL link. These t wo m essages are used for t he request and response of t he polling int erval of t he ACL link. < Day Day Up >

< Day Day Up >

Chapter 9. IEEE 802.15 The I EEE 802.15 group focuses on st andards t hat will cover Wireless Personal Area Net works ( WPANs) . So far I EEE 802.15 has int roduced one st andard, referred t o as I EEE 802.15.1, which st andardized part s of Bluet oot h. I n addit ion, t he group works on addit ional st andards t hat will include a high dat a rat e WPAN, referred t o as I EEE 802.15.3, and a low dat a rat e WPAN, referred t o as I EEE 802.15.4. The group is also developing recom m ended pract ices t o facilit at e t he coexist ence of I EEE 802.15 and I EEE 802.11, referred t o as I EEE 802.15.2. < Day Day Up >

< Day Day Up >

9.1 IEEE 802.15.1 I EEE 802.15.1 st andard adopt ed t he Bluet oot h Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions in early 2002. I t adopt ed Bluet oot h Version 1.1 specificat ions for lower t ransport layers ( L2CAP, Link Manager Prot ocol, and Baseband) and radio ( see Figure 9.1) .

Figu r e 9 .1 . M a ppin g of I SO OSI t o I EEE 8 0 2 .1 5 .1 W PAN St a n da r d

I EEE 802.15.1 PHY and MAC key m echanism s are alm ost ident ical t o t he m echanism s int roduced in t he Bluet oot h specificat ions. I EEE 802.15.1 specifies t he Logical Link Cont rol ( LLC) which support s t he LLC funct ion. Therefore, t he readers can refer t o appropriat e sect ions of Chapt er 8 for inform at ion regarding t he physical layer, Baseband, link m anager, Logical Link Cont rol and Adapt at ion Prot ocol ( L2CAP) , and Host Cont rol I nt erface ( HCI ) in I EEE 802.15.1. < Day Day Up >

< Day Day Up >

9.2 IEEE 802.15.3 The I EEE 802.15 group is working on a high dat a rat e WPAN t hat can support connect ivit y needs bet ween port able devices. This sect ion provides det ails on a pot ent ial fut ure st andard. However, t he fut ure st andard m ay be com plet ely different from t his descript ion. The working group is m ot ivat ed by t he need t o dist ribut e video and audio t hat requires m uch m ore bandwidt h t han I EEEE 802.15.1 can sust ain. The proposed radio is based on single carrier Quadrat ure Am plit ude Modulat ion ( QAM) wit h Trellis Coded Modulat ion ( TCM) working in t he 2.4 GHz band. Several select able speeds are discussed such as 11, 22, 33, 44, and 55 Mbps along wit h t hree t o four non- overlapping channels. I EEE is also considering anot her radio t echnology based on Ult ra Wide Band ( UWB) t hat will allow speeds of around 100 Mbps wit hin a 10 m et er range and 400 Mbps wit hin a 5 m et er range. I n addit ion t o dat a applicat ions, t hey envision applicat ions such as: high- speed t ransfer of digit al video from a digit al cam corder t o a TV or PC, int eract ive gam es t hat are m edia rich, hom e t heat er, and delivery of m ult im edia from a PC t o an LCD proj ect or. The MAC provides support for m ult im edia QoS requirem ent s via a TDMA based superfram e archit ect ure wit h Guarant eed Tim e Slot s ( GTSs) .

9.2.1 IEEE 802.15.3 Architecture 9.2.1.1 Network Topology The archit ect ure is based on t he concept of a piconet which is est ablished ad hoc and allows a num ber of independent dat a devices ( DEVs) t o est ablish peer- t o- peer com m unicat ion ( see Figure 9.2) . I t allows m obile devices t o j oin and leave t he piconet wit h short associat ion t im es. The piconet is confined t o a personal area and t ypically is expect ed t o cover a range of at least 10 m et ers and possibly up t o 70 m et ers. A piconet is est ablished once a DEV t hat is capable of becom ing t he Piconet Coordinat or ( PNC) begins t o t ransm it cont rol m essages called beacons. Through t he beacon, t he PNC m aint ains net work synchronizat ion by providing t im ing inform at ion. The PNC is also responsible for cont rolling t he channel access, m anaging t he QoS requirem ent s, perform ing adm ission cont rol, and assigning t im e slot s for connect ions bet ween DEVs. Before t ransm ission of dat a in a piconet , a DEV is required t o associat e wit h t he PNC and follow t he channel access inform at ion provided in t he beacons.

Figu r e 9 .2 . I EEE 8 0 2 .1 5 .3 Picon e t Ele m e n t s

Several piconet s can coexist by sharing t he sam e channel frequency. The piconet s are com bined t o cover a larger area and include m ore DEVs ( see Figure 9.3) . There is a proposal for a child piconet t o be est ablished for an exist ing piconet . Hence, t he child piconet can ext end t he area of coverage of t he exist ing piconet , increase t he num ber of DEVs t o be support ed, or shift som e t asks t o t he new piconet . When such a child piconet is est ablished, t he exist ing piconet becom es t he parent piconet . The child piconet uses a unique piconet I D ( PNI D) and funct ions alm ost independent ly. The child PNC handles associat ion, aut hent icat ion, securit y, and acknowledgm ent s wit hout t he parent PNC. The DEVs t hat associat e t o t he child piconet can com m unicat e wit h each ot her wit hin t he child piconet . However, t o avoid t he int erference bet ween t he parent piconet and child piconet s, t he parent PNC has t o allocat e GTSs or reserve t he t im e slot for t he child piconet t o access t he channel. DEVs of t he child piconet are allowed t o exchange dat a only wit hin t he allocat ed GTS. The child piconet 's PNC is also a m em ber of t he parent piconet and t hus can com m unicat e wit h t he DEVs in bot h t he parent piconet and it s own child piconet .

Figu r e 9 .3 . Pa r e n t / Ch ild/ N e igh bor Picon e t s

The com m it t ee also proposed t he idea of est ablishing a neighbor piconet under an exist ing piconet t hat will becom e t he parent piconet . This will provide t he m echanism for sharing t he frequency spect rum bet ween different piconet s when t here are no vacant physical radio channels. The neighbor piconet will use a unique PNI D. Sim ilar t o t he child piconet , it will funct ion alm ost independent ly from t he parent piconet . I t will handle associat ion, aut hent icat ion, securit y, and acknowledgm ent s wit hout involving t he parent piconet . However, t he parent piconet has t o allocat e GTSs for t he neighbor piconet t o access t he channel. Unlike t he PNC of t he child piconet , t he neighbor's PNC is not a m em ber of t he parent piconet . The neighbor's PNC is allowed t o send t o t he parent piconet only cert ain com m ands such as associat ion request , disassociat ion, channel t im e request ( CTR) , and aut hent icat ion. The creat ion process of child and neighbor piconet s is described in Sect ion 9.4.2.

9.2.1.2 Protocol Stack As shown in Figure 9.4, t he draft st andard divided t he prot ocol st ack int o various layers: fram e convergence sublayer ( FCSL) , MAC sublayer, PHY sublayer and layer- dependent m anagem ent ent it y ( MLME, PLME) . The fram e convergence sublayer provides fram e m ult iplexing from different prot ocols ( i.e., 802.2, 1394, USB, et c.) . The MAC sublayer defines t he following funct ionalit ies: associat ion/ disassociat ion, channel access, channel t im e m anagem ent , channel synchronizat ion, fragm ent at ion/ defragm ent at ion, acknowledgm ent / ret ransm ission, peer discovery, m ult irat e support , dynam ic channel select ion, and power m anagem ent . The PHY sublayer defines t he PHY specificat ions. The Device Managem ent Ent it y ( DME) which is a layerindependent ent it y provides t he st at us of various layer m anagem ent ent it ies. The st andard does not det ail t he DME funct ions, but rat her provides a general overview of it s funct ionalit y. Each ent it y int eract s t hrough a Service Access Point ( SAP) . Som e SAPs ( i.e., MAC SAP, PHY SAP, MLME SAP, MLME- PLME SAP) are explicit ly defined wit hin t he draft st andard but som e,

such as t he int erface bet ween t he MAC and t he MLME as well as t he int erface bet ween PHY and PLME, are not explicit ly defined.

Figu r e 9 .4 . I EEE 8 0 2 .1 5 .3 Pr ot ocol St a ck

< Day Day Up >

< Day Day Up >

9.3 IEEE 802.15.3 Physical Layer I EEE 802.15.3 operat es in t he 2.4 t o 2.4835 GHz frequency band. As shown in Table 9.1, t here are a t ot al of five channels in t wo set s of channel assignm ent s: high- densit y set and 802.11b coexist ence set . Each channel has 15 MHz bandwidt h. For t he high- densit y set ( 802.11b does not coexist ) four channels are defined while t hree channels are defined in 802.11b coexist ence set . I f a DEV det ect s 802.11b net work operat ing in it s area, it should use 802.11b coexist ence channel set . The PNC has t he abilit y t o select a channel for t he piconet operat ion. Using inform at ion such as int erference from ot her users, ot her I EEE 802.15.3 piconet s, and ot her unlicensed wireless ent it ies ( i.e., I EEE 802.11b) it can m ake a decision on t he best channel for it s piconet . The PNC has t he capabilit y t o dynam ically and t ransparent ly change t he channel on which t he piconet is operat ing wit hout requiring eit her user int ervent ion or int errupt ion of ongoing applicat ions.

Ta ble 9 .1 . 2 .4 GH z Ch a n n e l Assign m e n t Ch a n n e l I D Ce n t e r fr e qu e n cy

H igh - de n sit y ch a n n e l se t ( 8 0 2 .1 1 b doe s n ot coe x ist )

8 0 2 .1 1 b coe x ist e n ce ch a n n e l se t

X

1

2.412 GHz

X

2

2.428 GHz

X

3

2.437 GHz

4

2.445 GHz

X

5

2.461 GHz

X

X

X

The DEVs can discover inform at ion about t he services and capabilit ies of ot her DEVs in t he piconet . This inform at ion can be collect ed in several ways: t he PNC inform at ion request com m and, t he probe com m and, and t he piconet services inform at ion elem ent wit hin t he beacon. I n addit ion, t he PNC can ask a DEV in it s piconet t o evaluat e t he wireless m edia condit ions eit her in t he current channel or in an alt ernat e channel using t he rem ot e scan request com m and. A DEV can ask anot her DEV in t he piconet about t he st at us of t he current channel using t he channel st at us request com m and. The PNC inform at ion request com m and is used t o obt ain inform at ion from t he PNC about eit her a specific DEV or all of t he DEVs in t he piconet . The PNC reply includes t he I D and address of t he DEV, it s support ed dat a rat es, an indicat ion whet her t his DEV is PNC capable, an indicat ion whet her it is a PNC of anot her piconet , and ot her inform at ion. The PHY support s various dat a rat es corresponding t o t he m odulat ion and coding used such as 11 Mbps ( QPSK[ Quadrat ure Phase Shift Keying] - TCM) , 22 Mbps ( QPSK) , and 33, 44, 55 ( 16- 3264 QAM- TCM, respect ively) .

Recent ly, I EEE 802.15.3 est ablished t he 802.15.3a st udy group wit h t he chart er t o invest igat e alt ernat e PHY t o support very high dat a rat e applicat ions ( m ore t han 100 Mbps) . Exam ples of such applicat ions are: 1394a/ USB 2.0 high- speed cable replacem ent , high- densit y DVD, highresolut ion print er and scanner, and digit al cam era. Ult ra Wide Band ( UWB) is a pot ent ial PHY candidat e for such very high dat a rat e applicat ions. < Day Day Up >

< Day Day Up >

9.4 IEEE 802.15.3 Media Access Control I EEE 802.15.3 MAC includes several funct ions. I n t his subsect ion we will describe only som e key m echanism s such t he channel access and t he creat ion of a piconet .

9.4.1 Channel Definition and Channel Access The MAC is based on a t im e- slot t ed superfram e st ruct ure t hat consist s of t hree periods: Beacon, Cont ent ion Access Period ( CAP) , and Cont ent ion- Free Period ( CFP) ( see Figure 9.5) . Wit hin a CFP, t here are t wo t ypes of t im e slot s: m anagem ent t im e slot ( MTS) reserved for com m and exchange bet ween PNC and DEVs, and guarant eed t im e slot ( GTS) reserved for dat a exchange am ong DEVs. The lengt h of each period or t im e slot is det erm ined by t he PNC via beacons. At t he beginning of each superfram e, t he PNC broadcast s beacons t o specify t he cont rol, resource allocat ion, and t im e synchronizat ion t o t he ent ire piconet .

Figu r e 9 .5 . Su pe r fr a m e St r u ct u r e

9.4.1.1 Contention Access Period (CAP) I n t he CAP, cont ent ion is allowed via t he carrier sense m ult iple access wit h collision avoidance ( CSMA/ CA) m echanism . The DEVs can send sm all am ount s of pending dat a wit hout request ing reserved t im e slot s. Because of t he unpredict able channel access delay of t he cont ent ion schem e, CAP is suit able for asynchronous dat a t raffic such as file t ransfer, MP3 download files, et c. Moreover, CAP can be used for com m and exchange bet ween t he PNC and t he DEVs. CAP st art s aft er t he end of beacon and t erm inat es at a t im e specified by piconet synchronizat ion param et ers inform at ion elem ent in a beacon. Figure 9.6 shows t he CAP wit hin t he superfram e.

Figu r e 9 .6 . CAP St r u ct u r e in t h e Su pe r fr a m e

The draft st andard defines t wo int erfram e spacings ( I FSs) used in CSMA/ CA m echanism : short int erfram e space ( SI FS) and ret ransm ission int erfram e space ( RI FS) wit h RI FS = SI FS + aBackoffSlot . Bot h SI FS and aBackoffSlot durat ion are PHY dependent . The CSMA/ CA m echanism and backoff procedure are sim ilar t o t he I EEE 802.11 MAC int roduced in Chapt er 4 except for t he different I FSs ( i.e., RI FS is analogous t o DI FS in I EEE 802.11) . Figure 9.7 shows t he CSMA/ CA m echanism in CAP.

Figu r e 9 .7 . CSM A/ CA in CAP

A DEV t ransm it s a dat a packet aft er it senses t he channel idle for a t im e durat ion equal t o RI FS. Acknowledgm ent packet s are t ransm it t ed im m ediat ely aft er SI FS t im e. When a collision occurs ( i.e., m ult iple DEVs t ransm it packet s at t he sam e t im e) , t he exponent ial backoff algorit hm ( i.e., doubling t he cont ent ion window) is applied.

9.4.1.2 Contention-Free Period (CFP) The CFP is based on t im e division m ult iple access ( TDMA) in which t he PNC guarant ees t he st art ing t im e and reserved durat ion of t im e slot s for t raffic st ream s t hat require cert ain am ount s of bandwidt h. Tim e slot allocat ion is specified by t he channel t im e allocat ion ( CTA) inform at ion elem ent ( I E) in a beacon. As shown in Figure 9.8, t he CTA inform at ion elem ent includes source and dest inat ion addresses of t he DEVs t hat t ransm it and receive t he t raffic st ream , t he st ream index t hat ident ifies t he t raffic st ream , and t he st art ing t im e and durat ion of t he reserved t im e slot s ( MTS or GTS) .

Figu r e 9 .8 . Ch a n n e l Tim e Alloca t ion ( CTA)

Since t he bandwidt h is allocat ed using a TDMA schem e, t he channel is fully cont rolled by t he PNC and can provide QoS support . Therefore, CFP is suit able for isochronous dat a t raffic such as video st ream ing, audio st ream ing, and video/ audio broadcast . We will discuss QoS aspect s in Sect ion 9.5. As m ent ioned earlier, CFP includes t wo t ypes of t im e slot s: MTS and GTS. MTS is used for com m and exchange while GTS is used for dat a exchange. GTS can accom m odat e not only isochronous dat a t raffic but also asynchronous dat a t raffic. However, t he ways t o request and allocat e t im e slot s are slight ly different bet ween t he t wo t ypes of slot s. The PNC allocat es t im e slot s t hrough t he CTA m echanism . A source DEV, which has pending dat a t raffic and is seeking reserved t im e slot s for t ransm it t ing t he dat a, first sends t he channel t im e request ( CTR) com m and t o t he PNC t o indicat e t he recurring durat ion and t he num ber of required t im e slot s. For isochronous dat a t raffic, t he CTR com m and cont ains 1) t he num ber of t im e slot s needed ( i.e., t he num ber of t im e slot s per superfram e) and 2) t he durat ion of each t im e slot ( i.e., t he m inim um durat ion and desired durat ion) . I f t here are available resources or t im e slot s t o accom m odat e t he new request , t he PNC allocat es GTSs t o t he request ing DEV by including t hem in t he CTA inform at ion elem ent . On t he ot her hand, for asynchronous dat a t raffic, t he CTR com m and cont ains t he t ot al am ount of t im e needed t o t ransm it t he dat a inst ead of t he recurring channel t im e. The PNC will schedule such t im e based on t he channel availabilit y. The source DEV, dest inat ion DEV, and PNC are all allowed t o t erm inat e t he dat a connect ion whenever t hey want for bot h isochronous and asynchronous dat a t raffic. Each GTS is act ually one t im e slot reserved for a specific DEV. This DEV m ay or m ay not m ake use of all t he allocat ed t im e slot s. Each DEV is responsible for det erm ining how t o use it s reserved slot —t hat is, it can det erm ine what com m and, st ream , and asynchronous dat a will be t ransm it t ed. There are t wo t ypes of GTSs: dynam ic GTS and pseudost at ic GTS ( see Figure 9.9) . The t ype of GTS is also indicat ed in t he CTR com m and. The PNC m ay dynam ically change t he locat ion of t he dynam ic GTSs wit hin t he superfram e, on a superfram e- by- superfram e basis. This provides a t ool for t he PNC t o reassign GTS locat ions t o opt im ize channel ut ilizat ion. Pseudost at ic GTSs, which are allocat ed only for isochronous dat a t raffic, have fixed locat ion ( referenced t o t he beginning of t he beacon) wit hin t he CTA. The channel t im e locat ion of t he pseudost at ic GTS can be changed ( but less frequent ly t han dynam ic GTS) t o opt im ize t he channel ut ilizat ion. To carry out t his change, t he PNC has t o go t hrough a series of processes t o not ify t he involved DEVs ( bot h t ransm it t ing and receiving DEVs) . The PNC not ifies t he DEVs by sending a special com m and ( probe com m and) wit h t he new locat ion. Aft er all involved DEVs have been confirm ed, t he new CTA is included in t he beacon.

Figu r e 9 .9 . D yn a m ic GTS a n d Pse u dost a t ic GTS

Managem ent Tim e Slot s ( MTSs) are ident ical t o GTSs. MTSs are used for exchanging com m ands bet ween t he PNC and t he DEVs. There are t wo ways t o deploy MTSs: an MTS for t he com m and service of an individual DEV or an MTS for com m and service of m ult iple DEVs. I n t he lat t er case, t he MTS is denot ed as open MTS in which any DEV can t ransm it t he com m and m essages. The PNC det erm ines t he num ber of MTSs per superfram e. Since m ult iple DEVs can send com m and m essages in t he sam e MTS, collisions can occur. Therefore, in open MTS t he slot t ed Aloha prot ocol is used for channel access and cont ent ion resolut ion. Each DEV m aint ains it s own cont ent ion window ( CW) and a count er. The cont ent ion window is based on t he num ber of ret ransm ission at t em pt s of a DEV. The cont ent ion window is defined as follows:

Each DEV assigns t he count er value by random ly select ing an int eger in t he int erval [ 1,CW] . The DEV count s down in each open MTS cont inuously across superfram es. When t he count er reaches 1, t he DEV will t ransm it t he com m and m essage in t he current open MTS as shown in Figure 9.10. An acknowledgm ent ( ACK) is required t o indicat e t he success of t he com m and t ransm ission. An open MTS t hat is used for t he associat ion process is called associat ion MTS.

Figu r e 9 .1 0 . M TS Ch a n n e l Acce ss

9.4.2 Starting a Piconet The creat ion of a piconet begins wit h a DEV t hat is capable of being a PNC. This DEV scans t he channel and select s t he channel t hat has t he least int erference and is not being used by ot her piconet s. I f t he channel is available, t he DEV becom es a PNC and st art s sending beacons. Thus a piconet is est ablished. I f no channels are available, t he DEV can t ry t o est ablish a child or neighbor piconet inst ead.

9.4.2.1 Creating a Child Piconet A PNC capable DEV est ablishes a child piconet t o ext end t he coverage area of t he exist ing piconet , which event ually becom es t he parent piconet . The DEV sends t he channel request com m and t o t he parent 's PNC t o request t he special GTS ( called privat e GTS) for t he child piconet . I f a t im e slot is available, t he parent 's PNC will assign a privat e GTS t o t he child piconet . The DEV becom es t he child's PNC and st art s t ransm it t ing beacons. All com m unicat ion of t he child piconet t akes place wit hin t he assigned privat e GTS as shown in Figure 9.11. Because t he child's PNC is also a m em ber of t he parent piconet , t he parent 's PNC m ay allocat e GTS of t he parent piconet ( which is also in unassigned GTS durat ion of t he child piconet ) t o t he child's PNC t o enable int erpiconet com m unicat ion bet ween t he child's PNC and any DEVs ( including a PNC) of t he parent 's piconet .

Figu r e 9 .1 1 . Pa r e n t Picon e t a n d Ch ild/ N e igh bor Picon e t Su pe r fr a m e Re la t ion sh ip

9.4.2.2 Creating a Neighbor Piconet I f no channels are available during t he piconet creat ion procedure, a PNC capable DEV can est ablish a neighbor piconet on t he exist ing piconet , which event ually becom es t he parent piconet . First t he DEV sends an associat ion request t o t he parent 's PNC. I f t he neighbor associat ion request is accept ed, t he DEV sends t he channel request com m and t o t he parent 's PNC t o request privat e GTS for t he neighbor piconet . Then t he parent 's PNC will assign a privat e GTS t o t he neighbor piconet . The DEV becom es t he neighbor's PNC and st art s t ransm it t ing beacons. The superfram e relat ionship bet ween t he parent piconet and t he neighbor piconet is t he sam e as t he relat ionship bet ween t he parent piconet and t he child piconet as shown in Figure 9.11. Even t hough t he neighbor's PNC is not a m em ber of t he parent piconet , t he parent 's PNC m ay allocat e GTSs of t he parent piconet ( which is also in unassigned GTS durat ion of t he neighbor piconet ) . This enables t he neighbor's PNC t o have lim it ed com m unicat ion bet ween t he neighbor's PNC and t he parent 's PNC such as associat ion request s, disassociat ion request s, channel t im e request ( CTR) , and aut hent icat ion request s. < Day Day Up >

< Day Day Up >

9.5 IEEE 802.15.3 QoS Support I EEE 802.15.3 aim s t o support bot h asynchronous and isochronous dat a services. The draft st andard defines som e QoS m echanism s t o facilit at e t hose services. I EEE 802.15.3 MAC focuses on channel access m echanism s. I t leaves QoS negot iat ion and t raffic handling t o t he upper prot ocol layers. As we know, isochronous dat a is t im e sensit ive and requires cert ain bandwidt h and delay com m it m ent . As described in t he previous sect ion, I EEE 802.15.3 MAC has t wo m odes of operat ion: Cont ent ion Access Period ( CAP) and Cont ent ion- Free Period ( CFP) .

9.5.1 QoS in Contention Access Period A DEV wit h pending dat a is not required t o est ablish any connect ion or request any resource. A DEV accesses t he channel and com pet es for bandwidt h wit h ot her DEVs using t he CSMA/ CA m echanism . Therefore, CAP does not provide any service assurance—t hat is, it offers best effort service. CAP is suit able for asynchronous dat a services.

9.5.2 QoS in Contention-Free Period Using TDMA, t he PNC has full cont rol over t he channel by allowing a DEV t o access t he channel in a specific t im e for a specific durat ion. CFP provides cert ain QoS assurances in t erm s of bandwidt h and delay. I t aim s t o support high- bandwidt h asynchronous dat a and isochronous dat a. A priori t o t ransm ission, a connect ion needs t o be est ablished bet ween t he DEV and t he PNC. Such a connect ion is est ablished by sending t he PNC a channel t im e request ( CTR) com m and. I n case t here are available resources, t he PNC allocat es GTSs. The connect ion est ablishm ent process is shown in Figure 9.12.

Figu r e 9 .1 2 . Con n e ct ion Est a blish m e n t Pr oce ss

I n est ablishing isochronous connect ions, t he DEV m ust first det erm ine t he bandwidt h and delay requirem ent s of it s isochronous connect ion in t erm s of t he following param et ers: t he required num ber of t im e slot s ( GTSs) per superfram e, t he m inim um and desired durat ion of each t im e slot , priorit y ( based on eight priorit y values of I EEE 802.1p t raffic t ypes, see Table 9.2, and t he desired GTS t ype ( i.e., dynam ic GTS or pseudost at ic GTS) . These param et ers are included in t he CTR com m and. The m apping process bet ween t he QoS requirem ent s of an isochronous connect ion and t he CTR param et ers is t he t ask of each DEV and it is not specified by t he draft st andard. The PNC adm it s and allocat es GTSs based on t he CTR inform at ion.

Ta ble 9 .2 . I EEE 8 0 2 .1 p Pr ior it ie s a n d Tr a ffic Type s

Use r pr ior it y

Use d for

Com m e n t s

Best effort ( BE)

Asynchronous dat a

Default piconet t raffic

1

Background ( BK)

Asynchronous dat a

2

-

3

0 ( default )

Tr a ffic t ype

A spare

Current ly not assigned

Excellent effort ( EE)

I sochronous dat a

For valued cust om er

4

Cont rolled load ( CL)

I sochronous dat a

5

Video ( VI )

I sochronous dat a

< 100 m s delay and j it t er

6

Video ( VO)

I sochronous dat a

< 10 m s delay and j it t er

7

Net work cont rol ( NC)

I n est ablishing asynchronous connect ions, t he param et ers included in t he CTR com m and are slight ly different from t hose in t he isochronous connect ions. The DEV request s t he t ot al am ount of t im e for t he connect ion ( i.e., t ot al file t ransfer t im e) and t he priorit y. Aft er receiving t he CTR com m and, t he PNC perform s adm ission cont rol t o check if t here are available resources. When t here are available resources, t he PNC allocat es GTSs t o t he request ing DEV. The adm ission cont rol and scheduling algorit hm s are not specified in t he draft st andard.

9.5.2.1 Classification The draft st andard specifies t he st ream index used t o ident ify t he connect ion or t raffic st ream ( see Figure 9.13) . The st ream index is assigned by t he PNC during t he connect ion est ablishm ent procedure. There are t wo reserved st ream indices: 0x00 for unassigned st ream s and 0xFE for all asynchronous st ream s. A st ream index, except t hese t wo reserved st ream indices, is uniquely assigned for each isochronous st ream in t he piconet . The classificat ion process uses t he st ream index t o ident ify t he t raffic and forward it t o t he appropriat e queue. Due t o t he fact t hat t here is only a single st ream index for asynchronous st ream s, m ult iple asynchronous connect ions wit hin a DEV will be aggregat ed t o a single queue and wait for scheduling. Therefore, t here is no service different iat ion am ong asynchronous connect ions residing in t he sam e DEV. On t he ot her hand, each isochronous st ream is assigned a unique st ream index. This enables per- flow classificat ion.

Figu r e 9 .1 3 . Cla ssifica t ion

I n sum m ary, t he TDMA channel access schem e com bined wit h flow classificat ion will enable a cert ain level of QoS support for isochronous t raffic st ream s. However, t here are a num ber of QoS m echanism s ( i.e., adm ission cont rol, packet scheduling) which are st ill undefined. The net work designer m ay im plem ent per- flow QoS solut ions for which t he st andard defines som e necessary QoS m echanism s such as per- flow ( st ream index) classificat ion, CTA I E, and CTR com m and. The draft st andard also provides opt ional priorit y which m ay be used t o im plem ent different ial services. Figure 9.14 shows t he sim plified diagram of t he I EEE 802.15.3 QoS archit ect ure.

Figu r e 9 .1 4 . I EEE 8 0 2 .1 5 .3 QoS Ar ch it e ct u r e

< Day Day Up >

< Day Day Up >

9.6 IEEE 802.15.4 The I EEE 802.15.4 focuses on t he developm ent of a st andard t hat det ails t he wireless Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions for a Low- Rat e Wireless Personal Area Net work ( LR- WPAN) . The goal is t o int roduce a low- rat e WPAN t hat is sim ple and low- cost t hat support s wireless net work connect ivit y in applicat ions t hat require m inim al t hroughput as well as low power consum pt ion. I n t his sect ion we present som e of t he discussions and proposals in t he group regarding t he fut ure st andard. However, t he fut ure st andard m ay be com plet ely different . The group m ay also address locat ion t racking capabilit ies t hat can be applied wit hin sm art t ags and badges. Ot her applicat ions of t he I EEE 802.15.4 st andard m ay include personal net works, hom e net works, aut om ot ive net works, indust rial net works, int eract ive t oys, rem ot e sensing, and cable replacem ent for t he last m et er connect ivit y. Sensor net works m ay include hundreds or t housands of wireless sensors. Exam ples of sensor net works include wearable personal healt h m onit oring devices, hom e aut om at ion net works, and indust rial cont rol net works.

9.6.1 IEEE 802.15.4 Architecture 9.6.1.1 Network Topology LR- WPAN consist s of t wo or m ore int erconnect ed devices ( DEVs) t hat form a personal area net work, which som et im es is called a personal operat ing space ( POS) . I n t he draft st andard, t he DEVs are cat egorized int o t wo t ypes: a Full Funct ion Device ( FFD) and a Reduced Funct ion Device ( RFD) . The FFD can serve eit her as a net work coordinat or or as a sim ple net work node. A net work node ( which t ypically runs t he applicat ions) operat es as eit her a source or a dest inat ion of dat a t raffic, whereas t he net work coordinat or not only has t he net work node's funct ionalit y but also has t he abilit y t o rout e t he dat a t raffic, t o cont rol t he channel access of ot her devices, and t o provide basic t im ing wit hin t he POS. An FFD can com m unicat e wit h ot her RFDs or FFDs and can cont rol t he net work if it operat es as a net work coordinat or. An RFD can only be a net work node and can com m unicat e only wit h a net work coordinat or ( t here is no direct connect ion bet ween RFDs) . The RFD is geared for devices t hat are ext rem ely sim ple and do not need t o send large am ount s of dat a. Exam ples of such devices are light swit ches, act uat ors, sm art badges, and passive infrared sensors. DEVs can be organized in t hree net work t opologies: st ar t opology, peer- t o- peer t opology, and clust er t ree t opology ( see Figures 9.15 and 9.16) . FFDs can be part of any t opology; however RFDs are lim it ed t o a st ar t opology.

Figu r e 9 .1 5 . St a r a n d Pe e r - t o- Pe e r Topologie s

Figu r e 9 .1 6 . Clu st e r Tr e e Topology

I n a st ar t opology dat a m ay be exchanged only bet ween t he net work coordinat or and t he DEVs or net work nodes ( no direct com m unicat ion bet ween DEVs) . There are t wo t ypes of com m unicat ion: uplink, from t he DEVs t o t he net work coordinat or; and downlink, from t he net work coordinat or t o t he DEVs. I n a peer- t o- peer t opology, each DEV can com m unicat e wit h ot hers wit hin it s radio range. There is no net work coordinat or. However, a DEV can becom e such a coordinat or if it is t he first device t o com m unicat e on t he channel. A basic peer- t o- peer net work can be form ed when a new DEV searches for anot her DEV wit h which it can com m unicat e. I f no such DEVs are found, t he new DEV can becom e t he net work coordinat or and wait for ot her DEVs t o j oin. A m ore com plex t opology t hat can be const ruct ed out of t he peer- t o- peer t opology is a clust er t ree t opology ( Figure 9.16) . I n t he figure, t he circle area represent s a clust er. Mult iple clust ers int erconnect wit h each ot her, form ing a m ult i- clust er net work. Each clust er cont ains net work nodes and a net work node designat ed t o be a clust er head ( CH) . A clust er head cont rols it s clust er. All net work nodes in a clust er are only FFDs ( t here are no RFDs in a clust er t ree net work) . The clust er head of t he first est ablished clust er is called a designat ed device ( DD) . Therefore, t he ent ire m ult i- clust er net work will have only one DD. At t he clust er form at ion phase, a DD form s t he first clust er by assigning t he clust er ident ifier zero and t ransm it t ing t he beacon. A new net work node t hat is not associat ed wit h any clust er receives t he beacon and st art s m aking t he connect ion request t o t he clust er head ( or DD in t he case of t he first clust er) . The clust er head or DD will accept t he new net work node t o becom e a m em ber of t he clust er. Several net work nodes j oin t he clust er and t he clust er becom es bigger and bigger. The DD m ay prom ot e a net work node t o be t he clust er head and part it ion a big clust er int o sm all clust ers result ing in a m ult i- clust er net work. A unique clust er I D ( CI D) is assigned t o each clust er.

9.6.1.2 Protocol Stack

The proposed st andard is divided int o t he Link Layer Cont rol ( LLC) , MAC, and it s physical layer ( PHY) ( see Figure 9.17) .

Figu r e 9 .1 7 . I EEE 8 0 2 .1 5 .4 Pr ot ocol St a ck

< Day Day Up >

< Day Day Up >

9.7 IEEE 802.15.4 Physical Layer The radio can operat e at t he following frequencies: t he 868 MHz band ( e.g., in Europe) , t he 915 MHz band ( e.g., in t he U.S.) , and t he 2.4 GHz band ( worldwide) ( Figure 9.18) . At 868 MHz and 915 MHz, t he t ransm ission speed is 20 kbps using DSSS. 868 MHz band allows one channel while in t he 915 MHz band t here are 10 non- overlapping channels—t hat is, up t o 10 net works can coexist in t he sam e area. The 2.4 GHz band support s 250 kbps using DSSS, allowing 16 non- overlapping channels—t hat is, up t o 16 net works can coexist in t he sam e area.

Figu r e 9 .1 8 . I EEE 8 0 2 .1 5 .4 Fr e qu e n cy Ba n d

< Day Day Up >

< Day Day Up >

9.8 IEEE 802.15.4 Media Access Control I EEE 802.15.4 focuses on low- cost , low power consum pt ion devices. This focus const rains t he MAC prot ocol design t o be rat her sim ple wit h adequat e funct ionalit ies t o support t he low rat e applicat ions m ent ioned earlier. Sophist icat ed funct ions ( i.e., QoS m echanism s) reside in t he upper layers ( above I EEE 802.15.4 MAC) which are not included in t he draft st andard. Several t echniques are proposed t o m inim ize t he am ount of overhead in m aint aining t he com m unicat ion link. The packet st ruct ure is designed t o be sim ple. I EEE 802.15.4 has t hree packet st ruct ures: beacon packet , dat a packet , and handshake packet ( acknowledgm ent packet ) . Besides t he beacon, t here is no explicit m essage or com m and originat ed from t he MAC layer t hat provides peer com m unicat ion bet ween t he MAC layers of t he t wo devices. The draft st andard also defines several service prim it ives ( i.e., prim it ives bet ween t he LLC and MAC sublayers and prim it ives bet ween t he MAC sublayer and t he PHY) via t he service access point ( SAP) as shown in Figure 9.19. For m ore det ails on t he concept of service prim it ives, please refer t o I EEE 802.2. The MAC sublayer PAN inform at ion base ( MAC PI B) is t he MAC's dat abase which m aint ains t he net work configurat ion. The peer com m unicat ion bet ween t he MAC sublayers of t he t wo devices t akes place indirect ly t hrough t he upper layers and t he service prim it ives as shown in Figure 9.20.

Figu r e 9 .1 9 . M AC Su bla ye r Re fe r e n ce M ode l

Figu r e 9 .2 0 . Ex a m ple of Sign a lin g Flow

The proposed MAC ut ilizes eit her CSMA/ CA or TDMA prot ocols, or a com binat ion of bot h. As shown in Figure 9.21, t he channel is organized int o superfram es. Each superfram e st art s wit h a net work beacon, sent by t he net work coordinat or. The superfram e includes a beacon, a cont ent ion period, and guarant eed t im e slot s ( GTSs) . The superfram e durat ion varies depending on t he act ive applicat ions. I f t here are no low- lat ency applicat ions, t he superfram e consist s of only t he beacon and t he cont ent ion period. On t he ot her hand, if t here are lowlat ency applicat ions, GTSs are present .

Figu r e 9 .2 1 . I EEE 8 0 2 .1 5 .4 Fr a m e St r u ct u r e

I n t he beacon, t he net work coordinat or 1) synchronizes t he DEVs, 2) describes t he st ruct ure of t he superfram e, and 3) not ifies t he pending node m essages. I n t he cont ent ion period, t he DEVs can access t he wireless m edia using t he CSMA/ CA m echanism . The CSMA/ CA m echanism is sim ilar t o t he one present ed for t he I EEE 802.11 st andard ( see Chapt er 4) . Each t im e a DEV wishes t o t ransm it packet s, it needs t o det erm ine if t he wireless m edium is free. For t his process, t he MAC sends a clear channel assessm ent ( CCA) request t o t he physical layer. I f t he wireless m edium is free, t he packet is t ransm it t ed. I f t he wireless m edium is busy, t he DEV will back off for a random period before t rying again. All dat a t ransm issions require a handshake or acknowledgm ent by t he receiving DEV. I f an acknowledgm ent is not received at t he sender wit hin a pre- det erm ined t im e, t he sender will ret ransm it t he packet . The DEVs m ust st op com pet ing for channel access at t he end of t he cont ent ion period. The GTS is reserved for specific DEVs t hat need guarant eed bandwidt h. The net work coordinat or m ay decide not t o allocat e a GTS t o a request ing DEV or t o de- allocat e an exist ing GTS at any t im e. The proposal also defines t he following services required for LR- WPAN operat ion: net work discovery by t he net work coordinat or and DEV, net work init iat ion by t he net work coordinat or, net work synchronizat ion by t he DEV, and net work searching by t he DEV. Net work discovery is t he process of finding out which net works exist close t o t he net work adm inist rat or in order t o be able t o choose a unique net work ident ifier and channel. Net work init iat ion is t he process of est ablishing t he net work and it s operat ion. Net work synchronizat ion is t he process of list ening t o t he net work beacon t o find out t he availabilit y of com m unicat ion opport unit ies. Net work searching is t he process of finding t he net work beacon when net work synchronizat ion has been lost . < Day Day Up >

< Day Day Up >

9.9 IEEE 802.15.4 QoS Support I EEE 802.15.4 envisions t hree t raffic t ypes in t he LR- WPAN: ●

●

●

Periodic dat a: The t raffic is generat ed in a regular fashion. The am ount of dat a is defined by t he applicat ion it self and m ost ly is low dat a rat e. An exam ple is sensor t raffic. I nt erm it t ent dat a: The t raffic is generat ed once in a while, not cont inuously. The dat a generat ion is act ivat ed by ext ernal st im ulus. An exam ple is t he light swit ch t raffic. Repet it ive low- lat ency dat a: The t raffic is generat ed cont inuously and requires lowlat ency dat a t ransfer. An exam ple is t he m ouse device t raffic.

As described in t he previous sect ion, I EEE 802.15.4 includes t wo channel access schem es: CSMA/ CA and TDMA ( via GTS) . CSMA/ CA provides best effort service while TDMA provides quant it at ive service in t erm s of bandwidt h and delay assurance. Therefore, CSMA/ CA adequat ely support s periodic and int erm it t ent dat a while TDMA is applied t o repet it ive lowlat ency dat a. Repet it ive low- lat ency dat a t raffic requires GTSs from t he net work coordinat or. The net work coordinat or checks t he available net work resources. I f resources are available, t he net work coordinat or reserves GTSs as well as assigns unique I Ds t o each GTS. The negot iat ion process is accom plished t hrough t he upper layers and GTS prim it ives ( i.e., MLME- GTS.request , MLMEGTS.confirm ) . Figure 9.22 shows an exam ple of GTS prim it ives.

Figu r e 9 .2 2 . Ex a m ple s of GTS Pr im it ive s

Figure 9.23 shows t he GTS est ablishm ent procedure bet ween a net work node and t he net work coordinat or. The net work node's upper layer init iat es t he GTS request t o t he net work coordinat or. The upper layer of t he net work coordinat or ret rieves t he GTS request and signals t he MAC layer t hrough t he MLME- GTS.request . The MAC layer checks t he available net work resources. I f resources are available, t he request is grant ed, and t he MAC layer replies wit h MLME- GTS.confirm which cont ains t he grant ed GTS inform at ion ( i.e., GTS I D, GTS st art ing

t im e, GTS lengt h) . The net work coordinat or's upper layer includes grant ed GTS inform at ion in t he dat a payload and sends it t o t he net work node. MLME- GTS.confirm cont ains GTSI d, GTSSt art Slot , and GTSLengt h. GTSSt art Slot indicat es t he st art ing slot of t he allocat ed GTS. Aft er receiving t he grant ed GTS inform at ion, t he net work node's upper layer updat es it s MAC PI B wit h GTS inform at ion. The net work node accesses t he channel during a GTS based on t he inform at ion provided in t he MAC PI B. The upper layers' m echanism s are not specified by I EEE 802.15.4.

Figu r e 9 .2 3 . GTS Est a blish m e n t Pr oce du r e

I n sum m ary, I EEE 802.15.4 provides basic QoS m echanism s t o support low- rat e applicat ions. However, t he sophist icat ed QoS m echanism s and QoS algorit hm s reside in t he upper layer and are not wit hin t he scope of t he st andard. < Day Day Up >

< Day Day Up >

Part 5: 2.5G and 3G Networks

< Day Day Up >

< Day Day Up >

Chapter 10. GPRS Sect ion 10.1. I nt roduct ion Sect ion 10.2. GPRS ( Rel- 5) Archit ect ure Sect ion 10.3. Physical Channel Sect ion 10.4. Logical, Cont rol, and Traffic Channels Sect ion 10.5. Media Access Cont rol ( MAC) and Radio Link Cont rol ( RLC) Sect ion 10.6. Radio Resource Cont rol ( RRC) and Radio Resource ( RR) Sect ion 10.7. QoS Support < Day Day Up >

< Day Day Up >

10.1 Introduction General Packet Radio Service ( GPRS) is referred t o by m any as a 2.5G t echnology—an evolut ion from 2G Global Syst em for Mobile Com m unicat ions ( GSM) t echnology and an int erim phase t oward 3G high- speed services including m ult im edia t raffic wit h different QoS requirem ent s. GPRS st andardizat ion effort s were conduct ed by t he European Telecom m unicat ions St andards I nst it ut e ( ETSI ) during t he m id and lat e 1990s and t hen t ransferred t o t he 3G Part nership Proj ect ( 3GPP) organizat ion. 3GPP includes m any organizat ions such as t he Associat ion of Radio I ndust ries and Businesses ( ARI B) in Japan, t he China Wireless Telecom m unicat ion St andard Group ( CWTS) , t he European Telecom m unicat ion St andard Group ( ETSI ) , St andard Com m it t ee T1 Telecom m unicat ions ( T1) in t he U.S., and Telecom m unicat ion Technology Associat ion ( TTA) in Korea. ETSI , and lat er 3GPP, int roduced several releases of GPRS. I n each release new feat ures as well as increased speeds have been specified. Each release also has several versions in which problem s are clarified. Each release and version is det ailed in several docum ent s t hat provide t he release overview as well as det ailed specificat ions of it s various m odules and part s. 3GPP st art ed t o harm onize lat er releases of GPRS t o enable easy m igrat ion t o t he ant icipat ed 3G Universal Mobile Telecom m unicat ions Syst em ( UMTS) t echnology ( described in Chapt er 11) , which is provisioned by Wide Code Division Mult iple Access ( WCDMA) t echnology. The evolut ion of t hese st andards is present ed in Figure 10.1. I t is expect ed t hat m ost of t he t elephone carriers will t ake t his m igrat ion pat h, or part of it , t oward 3G. This includes bot h carriers t hat current ly support GSM and carriers t hat support alt ernat ive t echnologies such as CDMA. Such CDMA carriers are locat ed m ainly in t he U.S. and Korea.

Figu r e 1 0 .1 . Evolu t ion t ow a r d 3 G/ UM TS

GSM, which was int roduced in t he early 1990s, is available in m ore t han 100 count ries and is t he de fact o st andard in Asia and Europe. GSM is based on narrowband Tim e Division Mult iple

Access ( TDMA) , which can accom m odat e eight sim ult aneous calls on t he sam e frequency for circuit swit ching services. I t s st andards are specified by ETSI and t he TI A/ EI A ( Telecom m unicat ions I ndust ry Alliance/ Elect ronics I ndust ries Alliance) I S- 136 st andards. GSM charges are based on t he connect ion t im e and not on t he volum e of t raffic delivered. This cost m odel is not suit able for cert ain dat a applicat ions t hat are charact erized by burst y t ransm issions. GSM Phase 1 included com m on services such as basic t elephony, em ergency calls, up t o 9.6 kbps dat a rat e services, ciphering, aut hent icat ion, and feat ures such as call forwarding and SMS ( Short Message Service) . GSM Phase 2 included addit ional feat ures such as ident ificat ion, call wait ing, call hold, advice of charge, and m ult i- part y call. A half- rat e speech coding was int roduced in t his phase in addit ion t o full- rat e speech coding. These t wo phases provided t he basis for work t oward 3G, referred t o also as Phase 2+ , which includes GPRS specificat ions. The first work under Phase 2+ , referred t o as Release 96 ( Rel- 96) , specified HSCSD ( HighSpeed Circuit Swit ched Dat a) . HSCSD increased t he 9.6 kbps dat a rat e services provided in GSM Phases 1 and 2 t o a m axim um of 115 kbps. This dat a rat e is achieved t hrough a relat ively sim ple com binat ion of m ult iple GSM t im e slot allocat ions. However, due t o various im plem ent at ions and int erfaces, t he pract ical dat a rat e is up t o 64 kbps. I n HSCSD, t he dedicat ed t im e slot s t hat are reserved for dat a connect ions cause inefficient bandwidt h ut ilizat ion in cases where t he dat a t raffic is burst y. GPRS releases cont inued t o support HSCSD, which is current ly deployed only in a few locat ions. GPRS evolves on t op of GSM's TDMA approach. The first release of GPRS was det ailed in Release 97 ( R97) . R97 int roduced t he concept of packet swit ching in addit ion t o GSM's circuit swit ching. The packet swit ching allows carriers t o charge users based on act ual t raffic t ransm it t ed rat her t han on t he durat ion of t he circuit . GPRS also allows users t o access ot her public net works via prot ocols such as I P. R97 provides speeds of up t o 171 kbps wit h average speeds of 28.8 kbps. R97 allows support for QoS considering service priorit y, reliabilit y, t im e delay, t hroughput , QoS profile, and ot her param et ers. Release 98 ( R98) included Adapt ive Mult i Rat e ( AMR) t echnology, which provides t he abilit y t o select eit her full- rat e ( FR) or half- rat e ( HR) speech coding. These releases are being deployed globally despit e t heir relat ively low dat a rat es. Release 99 ( R99) includes Enhanced Dat a rat e for Global Evolut ion ( EDGE) , som et im es referred t o as Enhanced Dat a GSM Environm ent and as Enhanced GPRS ( EGPRS) . EDGE provided a num ber of addit ional radio m odulat ion and coding t echniques t hat enhanced speeds of up t o 384 kbps wit h average speeds of 64 kbps. This release includes EDGE Com pact , which is a variant of EGPRS t hat was designed t o be deployed in narrow frequency band allocat ions. EDGE Com pact requires less t han 1MHz of spect rum , com pared t o m ore t han 2.4 MHz for EDGE, reducing EDGE bandwidt h t o up t o 250 kbps wit h averages of up t o 56 kbps. R99 also included support for UMTS t raffic. This inclusion of UMTS t raffic has been cont inuously im proved and enhanced t hrough t he Release 4 ( Rel- 4) , t em porarily referred t o as Release 2000, and Release 5 ( Rel- 5) , known also as GSM/ EDGE Radio Access Net work ( GERAN) . I n GERAN a new radio access t echnology was int roduced t hat is fully harm onized wit h UMTS Terrest rial Radio Access Net work ( UTRAN) , which det ails speeds of up t o 2 Mbps. Rel- 5 includes an opt ional Enhanced GPRS ( EGPRS) radio t hat allows enhanced speeds com pared t o GPRS. Rel5 also includes a new set of m ult im edia services, referred t o as I nt ernet Mult im edia Subsyst em ( I MS) , including m ult im edia int eract ive com m unicat ions. 3GPP cont inues t o work on a newer release, referred t o as Release 6 ( Rel- 6) . Rel- 6 m ay include im proved radio t echnology for increased dat a rat es and ot her im provem ent s. I n t his book we focus on Rel- 5 while, when possible, we provide som e point ers t o previous releases. These

int erm ediat e releases receive int erest from various carriers t hat m ay deploy t hem as an int erm ediat e st ep t oward 3G t echnology. GPRS Release 5 is specified in several 3GPP docum ent s including: ●

●

●

●

●

●

●

●

●

●

TS 22.060 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Services and Syst em Aspect s; General Packet Radio Service ( GPRS) ; Service descript ion, St age 1 ( Release 5) TR 22.941 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Services and Syst em Aspect s; I P Based Mult im edia Services Fram ework; St age 0 ( Release 5) TS 22.228 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Services and Syst em Aspect s; Service requirem ent s for t he I P Mult im edia Core Net work Subsyst em ( St age 1) ( Release 5) TS 43.051 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; Overall descript ion ( St age 2) ( Release 5) TS 44.118 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM EDGE Radio Access Net work; Mobile radio int erface layer 3 specificat ion, Radio Resource Cont rol ( RRC) Prot ocol, I u Mode ( Release 5) TS 44.018 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; Mobile radio int erface layer 3 specificat ion; Radio Resource Cont rol Prot ocol ( Release 5) TS 44.060 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; General Packet Radio Service ( GPRS) ; Mobile St at ion ( MS) —Base St at ion Syst em ( BSS) int erface; Radio Link Cont rol/ Medium Access Cont rol ( RLC/ MAC) prot ocol ( Release 5) TS 44.160 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; Mobile St at ion ( MS) —Base St at ion Syst em ( BSS) int erface; Radio Link Cont rol/ Medium Access Cont rol ( RLC/ MAC) prot ocol; I u m ode ( Release 5) TS 43.064 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GERAN; Digit al cellular t elecom m unicat ions syst em ( Phase 2+ ) ; General Packet Radio Service ( GPRS) ; Overall descript ion of t he GPRS radio int erface ( St age 2) ( Release 5) TS 45.005 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; Radio t ransm ission and recept ion ( Release 5)

Several ot her docum ent s det ail various aspect s of t he st andards. They can be found at www.3gpp.org and www.et si.org. < Day Day Up >

< Day Day Up >

10.2 GPRS (Rel-5) Architecture As shown in Figure 10.2, t he GPRS m obile syst em consist s of t hree m ain ent it ies: Mobile St at ion ( MS) , Radio Access Net work ( RAN) , and Core Net work ( CN) . A m obile st at ion refers t o t he m obile equipm ent on which a user execut es m obile services and applicat ions ( i.e., voice services, SMS) . A m obile st at ion com m unicat es wit h t he base st at ion t hat cont rols t he radio channel wit hin it s coverage area ( or a cell, so t o speak) . Mult iple base st at ion syst em s int erconnect and form t he net work called RAN. RAN is responsible for providing t he com m unicat ion pat h bet ween t he Mobile St at ion and t he Core Net work. Furt herm ore, RAN also provides several key funct ions t o support com m unicat ion such as channel access, m obilit y m anagem ent , and radio resource m anagem ent . GERAN ( GSM/ EDGE Radio Access Net work) is t he RAN ent it y of GPRS ( Rel- 5) .

Figu r e 1 0 .2 . GPRS Ar ch it e ct u r e

The Core Net work connect s t he m obile net work t o t he ext ernal net work ( i.e., [ PSTN] , I nt egrat ed Services Digit al Net work [ I SDN] , I nt ernet ) . The core net work com prises t wo dom ains: Circuit Swit ching ( CS) dom ain and Packet Swit ching ( PS) dom ain. The CS dom ain

cont ains m odules t hat support circuit swit ching services. As shown in Figure 10.2, t he Mobile Swit ching Cent er ( MSC) resides in t he CS dom ain. The PS dom ain cont ains t he following m ain m odules t hat support packet swit ching services: Serving GPRS Support Node ( SGSN) and Gat eway GPRS Support Node ( GGSN) . A voice signal ( circuit swit ching t raffic) t ravels bet ween a m obile st at ion and an ext ernal voice net work ( i.e., PSTN) t hrough t he Base St at ion Syst em ( BSS) and MSC. On t he ot her hand, a dat a signal ( packet swit ching t raffic) t ravels bet ween a m obile st at ion and an ext ernal dat a net work ( i.e., I nt ernet ) t hrough t he BSS, SGSN, and GGSN. Hom e Subscriber Server ( HSS) is a support ing m odule providing subscript ion- relat ed inform at ion for bot h t he CS and PS dom ains. GERAN consist s of m ult iple int erconnect ed BSSs. Each BSS serves t he m obile st at ions wit hin it s coverage area. The BSS connect s t o t he CS dom ain of t he core net work t hrough A and I u- CS int erfaces while t he BSS connect s t o t he PS dom ain of t he core net work t hrough Gb and I u- PS int erfaces. I u- CS and I u- PS can be sim ply called I u int erface. The m obile st at ion connect s t o t he BSS t hrough t he Um int erface. Each int erface has it s own prot ocol archit ect ure t hat handles t he com m unicat ion bet ween t he pair syst em s. Prior t o Release 5, t he BSS connect s t o t he core net work only via t he A and Gb int erfaces ( no I u int erface) . This core net work is considered t o be t he second generat ion core net work. 3GPP realizes a sm oot h m igrat ion from t he second generat ion m obile syst em t oward 3G's UMTS ( Universal Mobile Telecom m unicat ions Syst em ) . I n UMTS, UTRAN ( UMTS Terrest rial Radio Access Net work) connect s t o t he t hird generat ion core net work t hrough t he I u int erface. Therefore, in order t o align wit h UMTS, GPRS Release 5 includes I u int erface in addit ion t o t he exist ing A/ Gb int erfaces. Figure 10.3 illust rat es t he GPRS archit ect ure for 2G and 3G core net works.

Figu r e 1 0 .3 . GPRS Ar ch it e ct u r e for 2 G a n d 3 G Syst e m s

The m obile st at ion can be in one of t wo m odes: A/ Gb m ode and I u m ode. The m ode is det erm ined by t he generat ion of t he core net work ( i.e., 2G or 3G) wit h which t he m obile st at ion associat es ( see Figure 10.3) . The st andard allows t hree dom ains of operat ion for t he m obile st at ion in A/ Gb m ode: ●

●

●

Class A dom ain of operat ion: The MS can operat e bot h packet swit ching services and circuit swit ching services sim ult aneously. Class B dom ain of operat ion: The MS can operat e eit her packet swit ching services or circuit swit ching service at one t im e. Class C dom ain of operat ion: The MS can only operat e packet swit ching services.

Furt herm ore, t he st andard also defines t hree dom ains of operat ion for t he m obile st at ion in I u m ode: ●

●

●

CS/ PS dom ain of operat ion: The MS can operat e bot h packet swit ching services and circuit swit ching services sim ult aneously. This dom ain is analogous t o Class A of t he A/ Gb m ode. PS dom ain of operat ion: The MS can only operat e packet swit ching services. This dom ain is analogous t o Class C of t he A/ Gb m ode CS dom ain of operat ion: The MS can only operat e circuit swit ching services.

GPRS support s bot h Point - t o- Point ( PTP) and Point - t o- Mult ipoint ( PTM) services while support ing roam ing of GPRS users bet ween different carriers. The PTP can be used for accessing inform at ion on t he I nt ernet , m essaging services, and conferencing applicat ions. PTM can be used for delivering inform at ion, such as news and weat her, t o m ult iple locat ions or int eract ive conferencing applicat ions. GPRS Rel- 5 added a new com ponent as com pared t o previous GPRS releases, t he I nt ernet Mult im edia Subsyst em ( I MS) . I MS allows sim ult aneous access t o m ult iple different t ypes of realt im e and non- real- t im e t raffic. Real- t im e t raffic includes voice, t ext , and video, and non- realt im e t raffic includes dat a, audio and video files, audio and voice st ream ing, t ext and m ult im edia m essaging, em ails, and I nt ernet browsing. I MS provides synchronizat ion bet ween t he m ult iple com ponent s of a m ult im edia com m unicat ion session. For exam ple, a user can receive calls while cont inuing wit h web browsing. I MS list s key requirem ent s for m ult im edia services. These requirem ent s pert ain t o t he user perspect ive. I n addit ion, I MS list s requirem ent s for securit y, addressing, support for roam ing, em ergency calls, m obile num ber port abilit y, m essaging, and Virt ual Privat e Net work ( VPN) support .

10.2.1 Radio Interface Protocol Architecture Figure 10.4 illust rat es t he BSS's radio int erface prot ocol st ack, which includes t he prot ocol m odules for I u and A/ Gb m odes. The prot ocol archit ect ure is divided int o 1) user plane for user dat a t ransfer and t he associat ed user dat a t ransfer cont rol and 2) cont rol plane for t he t ransfer of cont rol inform at ion t hat support s t he user plane.

Figu r e 1 0 .4 . BSS's Ra dio I n t e r fa ce Pr ot ocol St a ck

A logical channel t hat represent s a t raffic st ream of a specific t ype of inform at ion ( i.e., dat a m essage, cont rol m essage) cont ains a specific st ruct ure of t he inform at ion block ( or packet ) which is m apped t o t he physical layer. The logical channel will be described in a lat er sect ion. The radio int erface prot ocol is divided int o t hree layers: Layer 1 or physical layer ( PHY) ; Layer 2, which includes t he Dat a Link, Media Access Cont rol ( MAC) , Radio Link Cont rol ( RLC) and Packet Dat a Convergence Prot ocol ( PDCP) ; and Layer 3, which cont ains t he Radio Resource Cont rol ( RRC) for I u m ode and Radio Resource ( RR) for A/ Gb m ode.

10.2.1.1 RRC The Radio Resource Cont rol ( RRC) , int roduced in Rel- 5, is responsible for radio resource m anagem ent in t he I u m ode. I n t hese m at t ers it shares sim ilar responsibilit ies wit h t he MAC. The RRC will carry t he resource m anagem ent responsibilit ies when dedicat ed resources are provided t o t he m obile st at ion, where as t he MAC will be responsible when shared resources are provided t o t he m obile st at ion. The RRC broadcast s syst em inform at ion t o t he m obile st at ions. The syst em inform at ion m ay be cell specific ( such as locat ion services in cells) or not cell specific. The RRC is also responsible t o est ablish, m aint ain, reconfigure, and release connect ions bet ween t he m obile st at ions and t he GERAN. This includes cell reselect ion, adm ission cont rol, and link est ablishm ent . The RRC select s t he connect ion param et ers considering bot h cont rol and applicat ions needs. This allocat ion is t hen com m unicat ed t o t he m obile st at ion. The RRC also handles m obilit y relat ed act ivit ies such as evaluat ion, decision, and execut ion handovers. The RRC handover procedures follow t he m obilit y of t he m obile st at ion and include procedures t o m odify t he channels allocat ed t o t he m obile st at ion. While execut ing t hese funct ions, t he RRC considers t he request ed QoS. I n addit ion, t he RRC execut es a variet y of m onit oring, report ing, ciphering, and m ediat ing bet ween GERAN and t he m obile st at ions.

10.2.1.2 RR The Radio Resource ( RR) is responsible for t he resource m anagem ent in A/ Gb m ode. The RR is locat ed in a plane parallel t o t he RRC.

10.2.1.3 PDCP The Packet Dat a Convergence Prot ocol ( PDCP) is locat ed above t he RLC layer in t he user plane of t he I u m ode. PDCP includes t he following funct ions: ●

● ●

Header com pression and decom pression of t he I P dat a st ream : PDCP receives t he packet from t he upper layer, perform s header com pression, and forwards t he packet t o t he RLC layer. I n t he opposit e direct ion, PDCP receives t he packet from t he RLC layer, decom presses t he header, and forwards t he packet t o t he upper layer. Passage of t he dat a bet ween t he upper and lower layers. Maint enance of t he PDCP sequence num bers: PDCP assigns a sequence num ber t o t he packet t hat arrives from t he upper layer. At t he receiving end, PDCP receives t he packet wit h t he sequence num ber from t he RLC layer. Based on t he sequence num ber it decides it s act ion.

10.2.1.4 RLC The RLC t ransm it s dat a from t he higher layers t o t he MAC. I t can work in eit her t ransparent or non- t ransparent m ode. I n t ransparent m ode t he RLC has no funct ionalit y. Packet s are t ransferred t o t he MAC wit hout being alt ered and wit hout adding any RLC prot ocol inform at ion. I n non- t ransparent m ode, t he RLC provides reliabilit y t o t he MAC dat a t ransm ission. To achieve t his, t he RLC appends prot ocol inform at ion ( RLC packet header) t o t he dat a packet and perform s funct ions such as 1) segm ent at ion of upper layer packet s int o RLC dat a blocks, 2) reassem bly of RLC dat a blocks and t ransfer of t he blocks back t o t he upper layer, 3) concat enat ion of upper layer packet s int o RLC dat a blocks, 4) padding t o fill RLC dat a blocks, 5) backward Error Correct ion ( BEC) procedures enabling select ive ret ransm ission of RLC dat a blocks, 6) discarding of RLC packet s according t o delay requirem ent s, 7) in- sequence delivery of upper layer packet s, 8) Link Adapt at ion, 9) Ciphering, and 10) validat ing Acknowledgm ent s ( if desired by upper layers) . The non- t ransparent m ode can be divided int o Acknowledged m ode and Unacknowledged m ode. Acknowledgm ent m ode provides a m echanism for guarant eed delivery since t he addressee needs t o acknowledge t he receipt of t he packet . The RLC can not ify t he upper layers when errors cannot be resolved as well as when packet s are discarded. For unacknowledged m ode, RLC does not provide guarant eed delivery of packet s.

10.2.1.5 MAC The Medium Access Cont rol ( MAC) execut es t he funct ions of m anaging t he shared t ransm ission resources. I n t hese m at t ers it shares sim ilar responsibilit ies wit h t he RLC. The MAC responsibilit ies include 1) configuring and m apping logical channels, defined by t he t ype of dat a which is t ransferred, t o t he appropriat e physical subchannels; 2) assigning and configuring t he shared radio resources; and 3) scheduling, m ult iplexing dat a, and cont rolling packet s t o be delivered on t he shared physical channels. The MAC support s various logical Channels: t raffic channels ( TCH) , cont rol channels, and broadcast channels. The MAC support s bot h I u and Gb t raffic.

10.2.1.6 Data Link

The LAPDm delivers inform at ion bet ween Layer 3 and Layer 1 across t he GSM net work. The t erm Dm channel is used t o represent t he collect ion of all t he various signaling channels required in t he GSM syst em .

10.2.1.7 Physical Layer The physical channel int erfaces wit h t he Medium Access Cont rol ( MAC) and t he Radio Resource Cont rol ( RRC) . I t uses a com binat ion of frequency and Tim e Division Mult iple Access ( TDMA) m et hods. Each frequency channel is separat ed by 200 kHz. I n each frequency channel, t he channel is st ruct ured int o t im e slot s and defines eight basic physical channels per carrier. The physical channel is defined as a sequence of t im e slot s and a frequency hopping sequence. Several frequencies are defined for usage in different count ries. According t o 3GPP TS 45.001 docum ent , t he frequencies and channel arrangem ent s are as follows: 1. GSM 450 Band: t he syst em is required t o operat e in t he following band: a. 450.4 MHz t o 457.6 MHz: m obile t ransm it , base receive b. 460.4 MHz t o 467.6 MHz: base t ransm it , m obile receive 2. GSM 480 Band: t he syst em is required t o operat e in t he following band: a. 478.8 MHz t o 486 MHz: m obile t ransm it , base receive b. 488.8 MHz t o 496 MHz: base t ransm it , m obile receive 3. GSM 750 Band: t he syst em is required t o operat e in t he following band: a. 777 MHz t o 792 MHz: m obile t ransm it , base receive b. 747 MHz t o 762 MHz: base t ransm it , m obile receive 4. GSM 850 Band: t he syst em is required t o operat e in t he following band: a. 824 MHz t o 849 MHz: m obile t ransm it , base receive b. 869 MHz t o 894 MHz: base t ransm it , m obile receive 5. St andard or prim ary GSM 900 Band, P- GSM: t he syst em is required t o operat e in t he following frequency band: a. 890 MHz t o 915 MHz: m obile t ransm it , base receive b. 935 MHz t o 960 MHz: base t ransm it , m obile receive 6. Ext ended GSM 900 Band, E- GSM ( includes St andard GSM 900 band) : t he syst em is required t o operat e in t he following frequency band:

a. 880 MHz t o 915 MHz: m obile t ransm it , base receive b. 925 MHz t o 960 MHz: base t ransm it , m obile receive 7. Railways GSM 900 Band, R- GSM ( includes St andard and Ext ended GSM 900 Band) : t he syst em is required t o operat e in t he following frequency band: a. 876 MHz t o 915 MHz: m obile t ransm it , base receive b. 921 MHz t o 960 MHz: base t ransm it , m obile receive 8. DCS 1800 Band: t he syst em is required t o operat e in t he following band: a. 1710 MHz t o 1785 MHz: m obile t ransm it , base receive b. 1805 MHz t o 1880 MHz: base t ransm it , m obile receive 9. PCS 1900 Band: t he syst em is required t o operat e in t he following band: a. 1850 MHz t o 1910 MHz: m obile t ransm it , base receive b. 1930 MHz t o 1990 MHz base t ransm it , m obile receive To carry it s t ransm ission responsibilit ies successfully, t he physical layer is also responsible for power cont rol. The physical layer adj ust s t he m obile st at ion power out put level such t hat t he desired qualit y is achieved wit h t he m inim um possible power. I t synchronizes t he receiving st at ion wit h regard t o frequency and t im e. I t also carries handover funct ionalit y when a m obile st at ion m oves from one cell t o anot her as well as qualit y m onit oring t hat m ay result in a st at ion changing it s assigned physical channel. The channel m odulat ion can be eit her GMSK or 8- PSK. Each m odulat ion can achieve a different dat a rat e. The four channel coding schem es, CS- 1 t o CS- 4, are defined for GPRS, while nine channel coding schem es, MCS- 1 t o MCS- 9, are defined for EGPRS ( EDGE) . The bandwidt h allocat ed for speech or dat a t ransm ission is provided by a com binat ion of m odulat ion, block size, half or full rat e, code rat e, and ot her radio param et ers. The pot ent ial speed of speech, cont rol, and dat a varies significant ly. Speech varies up t o around 24 kbps, dat a varies up t o around 480 kbps.

10.2.2 GPRS (Rel-5) Protocol Architecture Figures 10.5 and 10.6 show t he user and cont rol plane prot ocols, respect ively, in t he PS dom ain ( bot h A/ Gb m ode and I u m ode) . Figures 10.7 and 10.8 show t he user and cont rol plane prot ocols, respect ively, in t he CS dom ain ( bot h A/ Gb m ode and I u m ode) . I n t hese figures we do not int end t o include all t he prot ocols' com ponent s and connect ions, but rat her t o provide t he big pict ure of t he GPRS ( Rel- 5) prot ocol archit ect ure.

Figu r e 1 0 .5 . Use r Pla n e Pr ot ocol Ar ch it e ct u r e in PS D om a in ( A/ Gb m ode a n d I u m ode )

Figu r e 1 0 .6 . Con t r ol Pla n e Pr ot ocol Ar ch it e ct u r e in PS D om a in ( A/ Gb m ode a n d I u m ode )

Figu r e 1 0 .7 . Use r Pla n e Pr ot ocol Ar ch it e ct u r e in CS D om a in ( A/ Gb m ode a n d I u m ode )

Figu r e 1 0 .8 . Con t r ol Pla n e Pr ot ocol Ar ch it e ct u r e in CS D om a in ( A/ Gb m ode a n d I u m ode )

As shown in Figure 10.5, t he funct ions bet ween A/ Gb m ode and I u m ode are slight ly different . I u RLC includes t he t ransparent m ode in addit ion t o t he acknowledged m ode and unacknowledged m ode in A/ Gb RLC. The PDCP used in I u m ode and t he Sub- Net work Dependent Convergence Prot ocol ( SNDCP) used in A/ Gb m ode provide t he logical link cont rol bet ween t he MS and t he BSS. I n Figure 10.5, t he RR operat es t he radio resource m anagem ent in A/ Gb m ode while t he RRC operat es t he radio resource m anagem ent in I u m ode. Ot her m odules include t he Radio Access Net work Applicat ion Part ( RANAP) , t he Signaling Connect ion Cont rol Part ( SCCP) , and t he Base St at ion Syst em GPRS Prot ocol ( BSSGP) . These m odules are not discussed in t his book. < Day Day Up >

< Day Day Up >

10.3 Physical Channel The physical channel is defined by a com binat ion of a radio frequency channel and a t im e slot num ber. Wit hin t he frequency band, a radio frequency channel ( seperat ed by 200 kHz) is st ruct ured int o t im e slot s ( TSs) as shown in Figure 10.9. Eight consecut ive t im e slot s form a TDMA fram e. A t im e slot in a TDMA fram e is num bered from 0 t o 7. A physical channel is defined by a radio frequency channel and a t im e slot num ber ( TN) in every TDMA fram e. All downlink and uplink TDMA fram es are aligned. At t he BSS, an uplink TDMA fram e is delayed by t hree t im e slot s from a downlink TDMA fram e. At t he m obile st at ion, t his delay is a funct ion of t he dist ance ( t he propagat ion delay) bet ween t he m obile st at ion and t he base st at ion.

Figu r e 1 0 .9 . Ph ysica l Ch a n n e l St r u ct u r e of St a n da r d GSM 9 0 0 Ba n d

A physical channel can be cat egorized as a Shared Basic Physical Sub Channel ( SBPSCH) or a Dedicat ed Basic Physical Sub Channel ( DBPSCH) . A DBPSCH is dedicat ed t o only one user. I n cont rast , SBPSCH can be shared by up t o eight users. Uplink St age Flag ( USF) is deployed t o cont rol t he m ult iple access of users on SBPSCH. We will describe t he USF m echanism in Sect ion 10.5.2. Figure 10.10 shows an exam ple of t he physical channel assignm ent .

Figu r e 1 0 .1 0 . Ex a m ple of Ph ysica l Ch a n n e l Assign m e n t

10.3.1 Hyperframes, Superframes, Multiframes, TDMA Frames and Time Slots As shown in Figure 10.11, t he st andard defines different durat ion fram es. The longest fram e, called hyperfram e, wit h a durat ion of 12533.76 seconds, is defined t o support crypt ographic m echanism s. A hyperfram e is divided int o 2048 superfram es, each wit h durat ion of 6.12 seconds. A superfram e is divided int o m ult ifram es. The num ber of m ult ifram es in a superfram e is based on t he t ype of m ult ifram es. There are four t ypes of m ult ifram es: ●

●

●

●

26- Mult ifram e: This is a 26- m ult ifram e t hat cont ains 26 TDMA fram es. Therefore, a superfram e can accom m odat e 51 26- m ult ifram es. 51- Mult ifram e: This is a 51- m ult ifram e t hat cont ains 51 TDMA fram es. Therefore, a superfram e can accom m odat e 26 51- m ult ifram es. 52- Mult ifram e: This 52- m ult ifram e, which is const ruct ed from t wo 26- m ult ifram es, cont ains 52 TDMA fram es. Therefore, a superfram e can accom m odat e 26 52m ult ifram es. 52- Mult ifram e for CTS ( Cordless Telephone Syst em ) : This m ult ifram e cont ains 52 TDMA fram es.

Figu r e 1 0 .1 1 . Tim e Fr a m e

Each t ype of m ult ifram e support s a different t ype of logical channel. Figure 10.11 shows only t he 26- m ult ifram e t ype. However, we will focus m ore on t he 52- m ult ifram e t ype, which is used t o accom m odat e dat a and cont rol packet s. A TDMA fram e cont ains eight t im e slot s, each wit h a durat ion of 0.577 m s. A t im e slot can accom m odat e a burst size of 156.25 sym bols. I f GMSK m odulat ion schem e is used, one sym bol reflect s one bit ( i.e., a t im e slot cont ains 156.25 bit s) . I f an 8- PSK m odulat ion schem e is used, one sym bol reflect s t hree bit s ( i.e., a t im e slot cont ains 469.75 bit s) .

10.3.2 Packet Data Channel (PDCH) The Packet Dat a Channel ( PDCH) refers t o t he physical channel dedicat ed t o packet dat a t raffic —t hat is, cont rol or user dat a t raffic from t he logical channel. PDCH uses 52- m ult ifram e st ruct ure. As shown in Figure 10.12, a PDCH m ult ifram e com prises 52 TDMA fram es which are divided int o: 12 blocks ( four TDMA fram es per block) , t wo idle fram es, and t wo fram es used for PTCCH ( Packet Tim ing Advance Cont rol Channel) .

Figu r e 1 0 .1 2 . 5 2 - M u lt ifr a m e for PD CH s

A PDCH is defined by a com binat ion of a radio frequency channel and a t im e slot . On a radio frequency channel, PDCHs can be represent ed logically as eight physical channel inst ances ( see Figure 10.13) . The PDCH num ber present ed here corresponds t o t he t im e slot num ber of t he TDMA fram e ( i.e., PDCH1 and t im e slot # 1)

Figu r e 1 0 .1 3 . PD CH s' Logica l Vie w

10.3.3 Radio Block The radio block is defined as a sequence of dat a burst s t hat cont ains one RLC/ MAC prot ocol dat a unit ( PDU) . A packet is generat ed at t he applicat ion layer and is sent t o t he lower layer. Each layer appends t o t he packet it s own prot ocol header and t railer. Figure 10.14 illust rat es t he packet flow and it s st ruct ure in each prot ocol layer. I t is wort h m ent ioning t hat t he RLC/ MAC PDU header cont ains t he Uplink St at e Flag ( USF) defined above, which plays an im port ant role in t he dynam ic channel assignm ent . At t he physical layer, t he RLC/ MAC PDU will be coded ( using a coding schem e such as CS- 1 t o CS- 4, or MCS- 1 t o MCS- 9) and reform at t ed t o fit int o a PDCH block.

Figu r e 1 0 .1 4 . Pa ck e t Flow a n d St r u ct u r e

< Day Day Up >

< Day Day Up >

10.4 Logical, Control, and Traffic Channels 10.4.1 Logical Channels The st andard defines logical channels t hat are m apped by t he MAC t o physical channels. These logical channels include bot h t raffic channels for dat a and speech and cont rol channels for cont rol, synchronizat ion, and signaling inform at ion. Several logical channels can be m ult iplexed on t he downlink physical channel. Sim ilarly, several logical channels can be m ult iplexed on t he uplink physical channel. Since GPRS is built on t he basis of GSM, som e logical channels are inherit ed from GSM. I n addit ion, GPRS int roduced som e new logical channels t o support packet swit ching services. As shown in Figure 10.15, t he nam ing of such logical channels begins wit h t he word " packet ," such as Packet Com m on Cont rol Channel ( PCCCH) and Packet Broadcast Cont rol Channel ( PBCCH) .

Figu r e 1 0 .1 5 . Pa ck e t D a t a Logica l Ch a n n e ls

10.4.2 Control Channels Cont rol channels carry cont rol, synchronizat ion, and signaling inform at ion. There are t hree kinds of cont rol channels: com m on, dedicat ed, and broadcast . Com m on cont rol channels include also Packet Com m on Cont rol Channels ( PCCCHs) t hat support packet t ransm ission. They include packet paging channels, random access channels, packet grant channels, and packet not ificat ion channels. PCCCH is opt ional for t he net work. I n case PCCCH is not allocat ed, t he cont rol inform at ion will be t ransm it t ed on t he GSM cont rol channels.

Dedicat ed cont rol channels support operat ions on DBPSCH. They include Slow Associat ed Cont rol Channel ( SACCH) for radio m easurem ent s and dat a. SACCH is also used for SMS t ransfer during calls. They also include Fast Associat ed Cont rol Channel ( FACCH) for one TCH on DBPSCH and St and- alone Dedicat ed Cont rol Channel ( SDCCH) . Cont rol channels support ing SBPSCH include cont rolling uplink and downlink act ivit ies using t he Packet Associat ed Cont rol Channel ( PACCH) , and Packet Tim ing Advance Cont rol Channels ( PTCCHs) for t im ing est im at ion. Broadcast channels include 1) frequency correct ion channels t hat are used t o correct t he frequency of t he m obile st at ion ( MS) , 2) synchronizat ion channels t hat synchronize t he MS frequency wit h t hat of t he base st at ion, 3) broadcast cont rol channels t hat broadcast general inform at ion on t he base st at ion, and 4) packet broadcast channels t hat broadcast param et ers t hat t he MS needs in order t o access t he net work for packet t ransm ission. The cont rol signaling m essages include t he following: 1. Packet Random Access Channel ( PRACH) and Com pact Packet Random Access Channel ( CPRACH) are used by t he m obile st at ion in t he uplink channel t o init iat e dat a or cont rol inform at ion. 2. Packet Paging Channel ( PPCH) and Com pact Packet Paging Channel ( CPPCH) are used t o page a m obile st at ion in t he downlink channel before dat a t ransm ission. 3. Packet Access Grant Channel ( PAGCH) and Com pact Packet Access Grant Channel ( CPAGCH) are used in t he downlink channel t o t ransm it t o t he m obile st at ion resource assignm ent s as a preparat ion st ep t oward dat a t ransm ission. 4. Packet Not ificat ion Channel ( PNCH) and Com pact Packet Not ificat ion Channel ( CPNCH) are used in t he downlink t o send a PTM- M ( Point t o Mult ipoint - Mult icast ) not ificat ion t o several m obile st at ions about a dat a t ransfer. 5. Packet Broadcast Cont rol Channel ( PBCCH) and Com pact Packet Broadcast Cont rol Channel ( CPBCCH) broadcast packet dat a specific syst em inform at ion in t he downlink. 6. Packet Associat ed Cont rol Channel ( PACCH) is used t o provide signaling inform at ion of a specified m obile st at ion such as acknowledgm ent s and power cont rol inform at ion. I t is also used t o t ransm it resource assignm ent and reassignm ent m essages. 7. Packet Tim ing Advance Cont rol Channel/ Uplink ( PTCCH/ U) is used t o t ransm it random access burst for est im at ion of t he t im ing advance for t he m obile st at ion in packet t ransfer m ode. 8. Packet Tim ing Advance Cont rol Channel/ Downlink ( PTCCH/ D) is used t o t ransm it t im ing advance inform at ion updat es t o several m obile st at ions. One PTCCH/ D is paired wit h several PTCCH/ Us. 9. Broadcast Cont rol Channel ( BCCH) in t he downlink is used t o broadcast cell specific inform at ion.

10.4.3 Packet Traffic Channels

Traffic channels ( TCHs) carry eit her encoded speech or user dat a. These channels are m ult iplexed in eit her a predet erm ined m anner or dynam ically by t he MAC. TCH can be eit her full rat e ( TCH/ F) or half rat e ( TCH/ H) . Traffic channels are also dist inguished by t he m odulat ion t echnique used such as GMSK and 8- PSK. Packet Dat a Traffic Channel ( PDTCH) is for dat a t ransfer on bot h physical channels SBPSCH ( Shared Basic Physical Sub Channel and DBPSCH ( Dedicat ed Basic Physical Sub Channel) . The PDTCH is a t em porary channel provided t o a single m obile st at ion or t o a group in case of m ult icast t ransm ission. However, one m obile st at ion m ay use several PDTCHs in parallel. PDTCH allows several MSs t o be m ult iplexed on t he sam e SBPSCH. PDTCH also allows several t raffic classes t o be m ult iplexed on t he sam e shared or dedicat ed channel. I n a m ult icast t ransm ission a m obile st at ion m ay use several PDTCHs for individual packet t ransfer. The PDTCH is eit her in t he uplink direct ion ( PDTCH/ U) or downlink direct ion ( PDTCH/ D) . A PDTCH m ay be eit her full rat e ( PDTCH/ F) or half rat e ( PDTCH/ H) depending on whet her it is t ransm it t ed on a Packet Dat a Channel/ Full ( PDCH/ F) or Packet Dat a Channel/ Half ( PDCH/ H) , respect ively. < Day Day Up >

< Day Day Up >

10.5 Media Access Control (MAC) and Radio Link Control (RLC) The Media Access Cont rol ( MAC) and Radio Link Cont rol ( RLC) provide dat a t raffic t ransfer from t he upper layer t o t he physical channel. The MAC is connect ion orient ed—t hat is, connect ions need t o be est ablished before any packet s can be t ransm it t ed. I n t he st andard such connect ions are called Tem porary Block Flows ( TBFs) . A TBF is a logical unidirect ional connect ion bet ween t wo MAC ent it ies, where one MAC ent it y belongs t o t he m obile st at ion and t he ot her t o t he BSS. The TBF is t em porary ent it y since t he connect ion ends once t here are no m ore dat a t o t ransfer and all acknowledgm ent s have been successfully received. The TBF is allocat ed radio resources on one or m ore physical channels ( i.e., PDCHs) . All packet s ( i.e., RLC/ MAC PDUs) t hat belong t o t he TBF are t ransm it t ed on t he allocat ed PDCHs. A PDCH can accom m odat e m ult iple TBFs. To different iat e bet ween TBFs, t he st andard int roduces Tem porary Flow I dent it y ( TFI ) . The TFI value is unique am ong concurrent TBFs in t he sam e direct ion ( uplink or downlink) . The TFI , t oget her wit h it s direct ion, provides a TBF unique ident ifier. I n addit ion, a Global_TFI is assigned t o each st at ion for ident ificat ion purposes. I n t he following subsect ions we will focus on som e key m echanism s incorporat ed in t he GPRS MAC.

10.5.1 TBF Establishment 10.5.1.1 TBF Establishment Initiated by Mobile Station The m obile st at ion which is t he source of t he dat a t raffic init iat es TBF est ablishm ent wit h t he BSS using one of t he following access processes: One Phase Access or Two Phase Access. These processes are described below.

10.5.1.1.1 One Phase Access I n t he One Phase Access ( Figure 10.16) , t he m obile st at ion init iat es t he packet access procedure by sending PACKET CHANNEL REQUEST m essages on t he PRACH channel. The PACKET CHANNEL REQUEST m essages cont ain an indicat ion of t he t ype of access, param et ers, and requirem ent s for radio resources. This includes t he request ed TBFs' t ype, Radio Priorit y, and num ber of radio blocks ( or RLC/ MAC blocks) . The num ber of request ed radio blocks cannot exceed t he act ual num ber of RLC/ MAC blocks wait ing in t he m obile st at ion. Due t o t he nat ure of t he random access schem e used in PRACH, it is possible t o have m ult iple st at ions sending PACKET CHANNEL REQUEST at t he sam e t im e ( i.e., collisions can occur) . I f a collision or packet loss occurs ( i.e., t he m obile st at ion does not receive a response from t he base st at ion) , t he PACKET CHANNEL REQUEST will be ret ransm it t ed. The m obile st at ion is allowed up t o a predet erm ined num ber of at t em pt s t o send a PACKET CHANNEL REQUEST. The m axim um num ber of at t em pt s is com bined wit h a persist ence level t hat is in direct relat ionship t o t he m obile st at ion priorit y. The higher t he priorit y, t he higher t he persist ence level. The num ber of allowed at t em pt s is broadcast ed by t he base st at ion on t he PBCCH and PCCCH.

Figu r e 1 0 .1 6 . TBF Est a blish m e n t I n it ia t e d by t h e M S: On e Ph a se Acce ss

Aft er successful recept ion of t he PACKET CHANNEL REQUEST, t he BSS replies on t he PAGCH channel wit h PACKET UPLI NK ASSI GNMENT. The PACKET UPLI NK ASSI GNMENT cont ains t he following inform at ion: ● ●

●

●

TFI : Unique value t hat ident ifies t he request ed TBF. One or m ore PDCHs assigned t o t he TBF: A radio frequency channel and one or m ore t im e slot s. USFs: 3- Bit field associat ed wit h each assigned PDCH. For exam ple, if a TBF connect ion is assigned frequency ci and t im e slot num bers 2 and 5, t here are t wo USFs assigned: one for t im e slot num ber 2 and t he ot her for slot num ber 5. USF_GRANULARI TY: 1- Bit field t hat indicat es t he num ber of RLC/ MAC blocks t ransm it t ed at a t im e. Value " 0" refers t o one RLC/ MAC block; value " 1" refers t o four consecut ive RLC/ MAC blocks.

10.5.1.1.2 Two Phase Access I n t he Two Phase Access ( Figure 10.17) , t he m obile st at ion init iat es t he packet access procedure by sending PACKET CHANNEL REQUEST. The BSS request s t he m obile st at ion t o send a PACKET RESOURCE REQUEST m essage. This request is sent im plicit ly t o t he m obile st at ion in t he PACKET UPLI NK ASSI GNMENT m essage. Upon receiving t he PACKET UPLI NK ASSI GNMENT, t he m obile st at ion t ransm it s a PACKET RESOURCE REQUEST m essage t o t he BSS on PACCH. The PACKET RESOURCE REQUEST cont ains t he following inform at ion: ● ●

Mobile st at ion radio access capabilit y. Channel Request Descript ion: t he request uplink resource includes t he radio peak t hroughput , priorit y class of t he request ed TBF, RLC m ode of t he request ed TBF ( i.e., RLC acknowledged m ode, RLC unacknowledged m ode) , and t he num ber of RLC/ MAC blocks t hat t he m obile st at ion wishes t o t ransm it .

Figu r e 1 0 .1 7 . TBF Est a blish m e n t I n it ia t e d by t h e M S: Tw o Ph a se Acce ss

Aft er receiving t he PACKET RESOURCE REQUEST, t he BSS replies on t he PACCH channel wit h PACKET UPLI NK ASSI GNMENT.

10.5.1.2 TBF Establishment Initiated by Network The net work m ay init iat e a TBF t o t ransfer packet s from t he net work t o t he m obile st at ion. The procedure m ay be ent ered when t he m obile st at ion is in MAC- I dle st at e ( for I u m ode) or in Packet I dle St at e ( for A/ Gb m ode) ( Figure 10.18) . Such TBF can also be init iat ed on PACCH if a TBF in t his direct ion is already est ablished. First t he BSS perform s a paging procedure t o discover t he locat ion of t he m obile st at ion. Aft er discovery, t he BSS will perform resource assignm ent . The net work m ay assign one or m ore PDCHs for t he TBFs based on t he net work's discret ion. The allocat ed radio resources are provided t o t he m obile st at ion in t he PACKET DOWNLI NK ASSI GNMENT m essage t ransm it t ed on t he PCCCH.

Figu r e 1 0 .1 8 . TBF Est a blish m e n t I n it ia t e d by N e t w or k

10.5.2 Channel Access and Resource Allocation The MAC is based on t he t im e- slot t ed st ruct ure of t he physical channel. There are t wo channel access m et hods: ●

●

Slot t ed Aloha m et hod: This random access m et hod is applied only in PRACH where m ult iple m obile st at ions send request packet s arbit rarily in t he uplink direct ion at t he beginning of a slot . I n case t he packet collides wit h a packet sent by anot her st at ion, t he st at ion backs off for an arbit rary t im e and t ries t o t ransm it again. Tim e Division Mult iple Access ( TDMA) m et hod: This cont ent ion- free channel access m et hod is applied in all logical channels except PRACH. The packet s are t ransm it t ed in predefined t im e slot s allocat ed by t he BSS. The st andard defines t hree t ypes of resource allocat ion m echanism s: Dynam ic Allocat ion, Ext ended Dynam ic Allocat ion, and Exclusive Allocat ion.

Using t he USF m echanism described in t he previous sect ions, t he net work can cont rol t he radio resource for each m obile st at ion dynam ically based on t he dem and of t he m obile st at ion ( as indicat ed in t he packet channel request ) and t he load of t he net work. USF is t he 3- bit field included in t he RLC/ MAC header of t he radio block ( Figure 10.14) sent by t he net work in t he downlink direct ion. This 3- bit field allows eight different USF values used for m ult iplexing in t he uplink channel. On PCCCH, one USF value ( USF = " 111" or called USF = FREE) is used t o indicat e PRACH. The ot her USF values are used t o reserve t he uplink for different MSs.

10.5.2.1 Uplink Data Transmission Aft er TBF est ablishm ent , a TFI and assigned PDCHs are associat ed wit h t he TBF. I n addit ion, a

USF is associat ed wit h an assigned PDCH. Up t o eight TBFs can be m ult iplexed on t he sam e PDCH. A m obile st at ion list ens and exam ines every RLC/ MAC PDU on t he assigned PDCHs ( in t he downlink) . I f t he m obile st at ion discovers it s USF in t he header of a RLC/ MAC PDU, t he following will occur. Depending on USF_GRANULARI TY ( assigned during t he TBF est ablishm ent procedure) t he MS will t ransm it st art ing at t he next uplink block, eit her one or four consecut ive radio blocks. Figure 10.19A shows an exam ple of t his procedure for t he Dynam ic Allocat ion m edium access m ode. The t able in Figure 10.19B includes t he TBFs at each MS. For TBF1 on MS1, USF GRANULARI TY= 0, result ing in one radio block t ransm ission in t he corresponding uplink block. I n t he case of TBF2 on MS2, USF GRANULARI TY= 1, result ing in four consecut ive radio block t ransm issions.

Figu r e 1 0 .1 9 . Ex a m ple s of USF M e ch a n ism s

Figure 10.19B shows an exam ple of t he USF procedure for t he Ext ended Dynam ic Allocat ion m edium access m ode where t here are m ult iple PDCHs assigned t o one connect ion. Not ice t hat in t his case if USF GRANULARI TY= 1, t he four radio blocks will be t ransm it t ed on different PDCHs.

10.5.2.2 Downlink Data Transmission For downlink t ransm issions which are governed only by t he BSS, t here is no need for t he USF m echanism . The BSS will choose t he appropriat e packet s for t ransm ission from t he TBF queues which reside at t he BSS. This t ransm ission will occur on t he downlink PDCHs associat ed wit h t he TBF.

10.5.3 Radio Link Control (RLC) The RLC layer, which is locat ed above t he MAC layer, can provide reliabilit y for MAC t ransm issions. The RLC can operat e in one of t he following t hree m odes: t ransparent m ode, acknowledged m ode, and unacknowledged m ode. The RLC in I u m ode includes all t hree m odes while RLC in A/ Gb m ode includes only acknowledged m ode and unacknowledged m ode. I n t ransparent m ode, t he RLC has no funct ionalit y—t hat is, t he packet s pass t hrough t he RLC layer t o t he MAC layer wit hout adding any prot ocol inform at ion. I n t he acknowledged m ode, t he t ransfer of packet s is cont rolled by a select ive Aut om at ic Repeat Request ( ARQ) m echanism , in which t he receiving st at ion request s ret ransm ission if an error occurs, coupled wit h t he num bering of packet s in one TBF. The sending st at ion t ransm it s packet s wit hin a window and t he receiving st at ion sends Packet Uplink Ack/ Nack or Packet Downlink Ack/ Nack m essages. The ARQ m echanism im plem ent ed is Select ive Repeat ARQ. I n ot her words, every Ack m essage acknowledges all correct ly received packet s up t o t he indicat ed sequence num ber. The receiving st at ion can select ively request erroneous packet s for ret ransm ission. The net work m ay reallocat e t he net work resources t o accom m odat e t he addit ional resources required for ret ransm issions. I n EGPRS, acknowledged m ode, t he t ransm ission of packet s is also cont rolled by m ore advanced ARQ m echanism s t hat enable t he ret ransm ission of only part of t he packet , reducing t he ret ransm issions overhead. I n unacknowledged m ode, t he t ransm ission of packet s is cont rolled by t he num bering of t he packet s wit hin one TBF and does not include any ret ransm issions. The receiving st at ion ext ract s user dat a from t he received packet s and at t em pt s t o preserve t he user inform at ion lengt h by replacing m issing packet s wit h dum m y inform at ion bit s.

10.5.4 MAC States The MAC as applied t o GERAN I u m ode can be in one of four st at es: MAC- I dle, MAC- Shared, MAC- DTM, and MAC- Dedicat ed. I n t he MAC- I dle st at e t here are no TBFs and t he m obile st at ion m onit ors relevant paging subchannels on t he PCCCH. Once t he upper layers have a packet for t ransm ission and connect

wit h t he MAC, t he MAC will est ablish a TBF on a shared channel, SBPSCH, and it t ransit ions t o MAC- Shared st at e. The RRC or RLC/ MAC m ay also dict at e t o t he MAC t o est ablish a TBF on a dedicat ed channel, DBPSCH. I n t his case t he MAC t ransit s t o a MAC- Dedicat ed st at e. I n t he MAC- Shared st at e, t he m obile st at ion is allocat ed radio resources providing a TBF for a point - t o- point , unidirect ional t ransfer of upper layer packet s on one or m ore shared channels ( SBPSCHs) . The MAC can t ransm it t hese packet s in eit her RLC acknowledged m ode or RLC unacknowledged m ode. When all packet s have been delivered and t he acknowledgm ent s have been received, t hen t he MAC can release all TBFs on t he downlink and uplink direct ion and ret urn t o t he MAC- I dle st at e. I n t he MAC- Dedicat ed st at e, t he m obile st at ion is allocat ed radio resources t o t ransm it a TBF on one or m ore dedicat ed channels ( DBPSCHs) . The MAC can t ransm it t hese packet s in RLC acknowledged m ode, RLC unacknowledged m ode, or t ransparent m ode. Aft er all packet s have been delivered and acknowledgm ent s have been received, t he MAC can release all TBFs on t he DBPSCHs and ret urn t o t he MAC- I dle st at e. I n t he MAC- Shared st at e, upper layers m ay require t he t ransfer of an upper layer packet , which m ay t rigger t he est ablishm ent of a TBF on a dedicat ed channel, DBPSCH. I n t his case t he MAC will t ransm it from it s MAC- Shared st at e t o t he MAC- DTM st at e. I n t he MAC- DTM st at e a m obile st at ion has been allocat ed bot h shared and dedicat ed radio channels ( one or m ore DBPSCHs and one or m ore SBPSCHs) . Packet s can be delivered in RLC acknowledged m ode, RLC unacknowledged m ode, or RLC t ransparent m ode. When all packet s t hat t riggered t he allocat ions of shared resources have been t ransm it t ed and consequent TBFs on t he downlink and uplink SBPSCHs have been released, t he MAC will t ransit t o MACDedicat ed st at e. Sim ilarly, when all packet s t hat t riggered t he allocat ion of dedicat ed channels have been t ransm it t ed and t heir consequent TBFs have been released t o t he MAC, t he MAC will t ransit t o MAC- Shared sat e. Once all TBFs on bot h SBPSCHs and DBPSCHs have been released, t he m obile st at ion ent ers t he MAC- I dle st at e. When t he MAC is in eit her MAC- I dle st at e or MAC- Shared st at e it cont inuously m onit ors t he syst em inform at ion broadcast ed in t he PCCCH. The MAC on t he m obile st at ion m ay indicat e t o t he RRC t he availabilit y of a new cell and a cell change. When a cell change is required eit her by t he m obile st at ion or by t he net work, t he m obile st at ion can operat e in MAC- I dle st at e or MAC- Shared st at e in t he current cell while get t ing syst em inform at ion on t he new cell. However, t he m obile st at ion m ay suspend it s TBF( s) and m onit or t he PCCCH in t he current cell, in order t o receive t he necessary inform at ion on t he new cell BCCH. Once t he swit ch t o t he new cell has been com plet ed, t he m obile st at ion abort s any TBF in progress on bot h t he downlink and uplink. The m obile st at ion will reest ablish it s required TBF connect ions in t he new cell. < Day Day Up >

< Day Day Up >

10.6 Radio Resource Control (RRC) and Radio Resource (RR) The Radio Resource Cont rol ( RRC) is responsible for radio resource m anagem ent for I u t raffic, whereas t he Radio Resource ( RR) is responsible for m anaging t he A/ Gb t raffic. The relat ionship bet ween t he t wo is present ed in Figure 10.20.

Figu r e 1 0 .2 0 . Re la t ion sh ip be t w e e n RR a n d RRC

The RRC and RR are responsible for allocat ing new dedicat ed basic physical subchannels as well as t he int racell handovers of t he dedicat ed basic physical subchannels. The RRC broadcast s syst em inform at ion ( such as locat ion services in cells) t o t he m obile st at ions. The RRC also est ablishes, m aint ains, reconfigures, and releases connect ions bet ween t he m obile st at ions and t he GERAN net work. This includes cell reselect ion, adm ission cont rol, and link est ablishm ent . This also includes select ion of t he connect ion param et ers considering bot h cont rol and applicat ions needs. This allocat ion is t hen indicat ed t o t he m obile st at ion. The RRC also handles m obilit y, which includes evaluat ion, decision, and execut ion of funct ions such as handover and m oving from cell t o cell. The RRC handover procedures cont rol t he m obilit y of t he m obile st at ion and include procedures t o m odify t he channels allocat ed t o t he m obile st at ion. Execut ing t hese funct ions, t he RRC considers t he request ed QoS and ensures allocat ion of sufficient resources for t he t arget ed QoS. I n addit ion, t he RRC perform s m onit oring, report ing, ciphering, and m ediat ing bet ween GERAN and t he m obile st at ions. The RR m aint ains at least one PDCH t hat carries user dat a and all t he necessary cont rol signaling for init iat ing packet t ransfer whenever t hat signaling is not carried by t he exist ing cont rol channels. Ot her PDCHs, act ing as slaves, are used for user dat a t ransfer and for dedicat ed signaling. The GPRS net work is based on dynam ically allocat ing capacit y and hence does not require perm anent ly allocat ed PDCHs. Capacit y allocat ion is based on act ual needs. A m obile st at ion can be allocat ed perm anent or t em porary physical resources ( i.e., PDCHs) . I n cases where som e PDCHs are congest ed, t he net work m ay allocat e m ore resources. The exist ence of PDCHs does not im ply t he exist ence of PCCCH. I n t his case, t he GPRS st at ions can

use t he GSM cont rol channel CCCH. The net work can t hen assign resources on PDCHs for uplink t ransfer. Aft er t he t ransfer, t he MS ret urns t o CCCH. However, when PCCCH is allocat ed in t he cell, all GPRS st at ions will use it . The net work can allocat e a PCCCH eit her as a result of t he increased dem and for packet dat a t ransfers or if t here are enough available physical channels. I f t he net work finds t hat t he PCCCH capacit y is not adequat e, it can allocat e addit ional PCCCH resources on one or several PDCHs. The RRC has several m odes of operat ion. When t he m obile st at ions powers on and t he I u m ode is select ed, t he MS ent ers t he RRC- I dle Mode. I n t his m ode t he MS m onit ors cont rol broadcast s m essages. Once a connect ion is est ablished and t he I u m ode is ent ered, t he RR- I dle m ode changes t o RRC- Connect ed m ode. Such a connect ion can be m ade only aft er t he upper layers have request ed a connect ion and t he net work has responded by assigning com m unicat ion resources. The RRC- Connect ed m ode is charact erized by t hree st at es: RRC- Cell_Shared, RRCCell_Dedicat ed, and RRC- GRA_PCH. The RR support s bot h GPRS packet t raffic and GSM t raffic circuit swit ched t raffic. The circuit swit ch part includes an I dle m ode and Dedicat ed m ode. The GPRS part includes Packet I dle m ode and Packet Transfer m ode. St at ions belonging t o Class A, support ing bot h GPRS and GSM circuit swit ching t raffic, are in Dual Transfer m ode.

10.6.1 RRC Modes of Operation The RRC has several m odes of operat ion: The RRC- Cell_Shared, RRC- Cell_Dedicat ed, and RRCGRA_PCH. I n RRC- Cell_Shared st at e t he MS execut es a cell updat e procedure on cell changes, m onit ors t he PBCCH cont rol channel for syst em inform at ion m essages, and m onit ors neighboring cells for neighbour cell m easurem ent s. The RRC will t ransit t o RRC- I dle m ode when t he RRC connect ion is released or when t he operat ion m ode is changed t o A/ Gb m ode. I n t he RRC- Cell_Shared st at e t he MAC is responsible for allocat ing t he shared physical subchannels. The RRC- Cell_Dedicat ed st at e is assigned when a Dedicat ed Basic Physical Sub Channel ( DBPSCH) is allocat ed t o t he MS. I n t his m ode t he MS is assigned one or m ore dedicat ed basic physical subchannels in t he uplink and downlink. A st at ion in t his st at e perform s m easurem ent and report ing act ivit ies, list ens t o neighboring cells for m easurem ent s, and execut es handover procedures of t he dedicat ed basic physical subchannels on a cell change. Transit ion from RRCCell_Dedicat ed st at e t o RRC- I dle m ode occurs when t he RRC connect ion is released. I n t he RRCCell_Dedicat ed st at e t he RRC is responsible for allocat ing t he physical dedicat ed subchannels. The RRC- Cell_Dedicat ed st at e changes t o RRC- Cell_Shared st at e when 1) all t he dedicat ed basic physical subchannels are released and shared basic physical subchannels exist or 2) no shared basic physical subchannels exist . The RRC- GRA_PCH st at e is assigned when GERAN orders t he MS t o m ove t o RRC- GRA_PCH st at e via explicit signaling. I n t his t ransit ion t he MS abort s any TBF in progress. I n RRCGRA_PCH st at e, no basic physical subchannel is allocat ed t o t he MS and no radio resource allocat ion t asks are execut ed. Hence, no uplink act ivit y is possible. However, t he MS cont inues t o m onit or t he cont rol channel; PCCCH and it s locat ion is known t o GERAN. RRC will change from RRC- GRA_PCH st at e t o RRC- Cell_Shared st at e due t o changes in GRA updat e, cell updat e, or response t o paging. Also in t he RRC- GRA_PCH st at e, t he MS m ay request a radio resource t o answer t o a paging m essage or t o perform a GRA/ Cell updat e procedure.

The RRC in RRC- Connect ed m ode is responsible for allocat ing dedicat ed basic physical subchannels, which causes t he MS t o ent er t he RRC- Cell_Dedicat ed st at e. The MAC is responsible for allocat ing shared basic physical subchannels ( SBPSCH) . The MAC allocat es t he PDTCHs according t o t he QoS class of t he radio link and t he m ult i- slot capabilit y of t he MS. The RRC provides t he MAC wit h QoS class and indicat ion of t he MS m ult i- slot capabilit y.

10.6.2 RR Modes of Operation The RR support s bot h GPRS packet swit ched t raffic and GSM circuit swit ched t raffic. The circuit swit ched part includes an idle m ode and a dedicat ed m ode as well as ot her m odes relat ed t o voice services t hat are not discussed in t his book. The GPRS part includes packet idle m ode and packet t ransfer m ode.

10.6.2.1 Packet Idle Mode I n packet idle m ode, t he MS list ens t o t he PBCCH and t o t he paging subchannel for t he paging m essages. I f PCCCH is not present in t he cell, t he m obile st at ion list ens t o t he BCCH and t o t he relevant paging subchannels. A st at ion in t he packet idle m ode t hat belongs t o GPRS class A m ay sim ult aneously ent er t he different RR service m odes defined for GSM circuit swit ch services ( e.g., circuit swit ched connect ions m anagem ent , m obilit y m anagem ent , and radio resource m anagem ent ) . I n t his m ode upper layers can require t he t ransfer of packet s t hat result s in t he est ablishm ent of TBFs and t ransit ion t o packet t ransfer m ode. This m ode is not applicable t o an MS support ing DTM wit h an act ive RR connect ion.

10.6.2.2 Dedicated Mode An MS t hat support s DTM, m eaning support ing bot h GPRS and ot her GSM services sim ult aneously, and has an act ive RR connect ion and has no allocat ed packet resources is in dedicat ed m ode.

10.6.2.3 Packet Transfer Mode I n packet t ransfer m ode, t he m obile st at ion is allocat ed radio resources for it s TBFs on one or m ore physical channels on bot h t he uplink and downlink. Delivery of t he packet s can be done in RLC acknowledged or RLC unacknowledged m odes. Packet t ransfer m ode is not applicable t o a m obile st at ion support ing DTM t hat has an act ive RR connect ion. When select ing a new cell, t he m obile st at ion leaves t he packet t ransfer m ode and ent ers t he packet idle m ode. Aft er m oving t o t he new cell, it reads t he syst em inform at ion and m ay change back t o Packet Transfer Mode and resum es packet t ransm ission operat ions. A GPRS class A st at ion m ay sim ult aneously ent er different RR service m odes for GSM circuit swit ched services.

10.6.2.4 Dual Transfer Mode (DTM) I n dual t ransfer m ode t he m obile st at ion support s an act ive RR connect ion and is allocat ed radio resources for it s TBFs on one or m ore physical channels in t he uplink and downlink direct ions. This m ode is only applicable for a m obile st at ion support ing GPRS or EGPRS and is a subset of class A m ode of operat ion ( i.e., support ing bot h GPRS and ot her GSM services sim ult aneously) .

Packet s can be t ransferred in RLC acknowledged or RLC unacknowledged m ode. A st at ion in DTM is carrying out all of t he t asks of a st at ion in t he dedicat ed m ode. I n addit ion, t he upper layers can require t he release of all t he packet resources, which t riggers t he t ransit ion t o t he dedicat ed m ode, and t he release of t he RR resources, which t riggers t he t ransit ion t o idle m ode and packet idle m ode. When t he st at ion is handed over t o a new cell, t he RR leaves t he DTM, ent ers t he dedicat ed m ode, m ay read t he syst em inform at ion m essages sent on t he cont rol channels, and t hen ent ers t he DTM.

10.6.3 Transition between Modes Figure 10.21 shows t he four RR st at es for a class A st at ion t hat does not support DTM. The four st at es can be regarded as a com binat ion of t wo st at e m achines wit h t wo RR st at es each: circuit swit ched part com prises idle m ode and dedicat ed m ode and t he GPRS part com prises packet idle m ode and packet t ransfer m ode.

Figu r e 1 0 .2 1 . RR M ode s for a Cla ss A t h a t D oe s N ot Su ppor t D TM

Figure 10.22 shows t he RR m odes and t ransit ions for a class A st at ion t hat support s DTM and

class B. Class B includes t he following m odes: packet idle m ode, packet t ransfer m ode, and dedicat ed m ode.

Figu r e 1 0 .2 2 . RR M ode s a n d Tr a n sit ion s for Cla ss B a n d Cla ss A t h a t Su ppor t s D TM

Figure 10.23 shows t he RR m odes and t ransit ions for a Class C st at ion. When it is at t ached t o a GSM channel, t here are t wo RR m odes: idle m ode and dedicat ed m ode. When it is at t ached t o a GPRS channel, t here are t wo RR m odes: packet idle m ode and packet t ransfer m ode.

Figu r e 1 0 .2 3 . RR M ode s a n d Tr a n sit ion s for Cla ss C

< Day Day Up >

< Day Day Up >

10.7 QoS Support The st andard realizes t hat successful deploym ent of I P m ult im edia services requires support and m anagem ent capabilit y t o provide different degrees of end- t o- end QoS under various condit ions. End- t o- end QoS m ay be specified by Service Level Agreem ent s. Hence, it requires considerat ion of available bandwidt h, QoS classes, m easurem ent , m onit oring, and ot her t ools. I t requires t hat QoS support for I P m ult im edia sessions includes t he following funct ions: ●

●

●

● ● ● ● ●

Negot iat e QoS for I P m ult im edia sessions as a whole as well as t heir individual com ponent s, when t he session is est ablished and during it s execut ion. The assum pt ion is t hat t he I P m ult im edia applicat ions are able t o define t heir requirem ent s, negot iat e t heir capabilit ies, and ident ify and select t he available m edia com ponent s, such as QoS. Allow eit her t he user or t he net work t o ident ify alt ernat ive dest inat ions for I P m ult im edia sessions and t heir individual com ponent s. Based on t his ident ificat ion, t hese sessions can be redirect ed t o t hese alt ernat e dest inat ions at various st ages of t he m ult im edia sessions. Support end- t o- end QoS for voice t hat has qualit y at least as good as t hat provided by circuit - swit ched net works. Support roam ing and negot iat ion bet ween carriers for QoS. Deploy I P Policy Cont rol for I P m ult im edia applicat ions. Support m ore t han one I P m ult im edia applicat ion in a session. Provide securit y services sim ilar t o t hose of circuit swit ched net works. Support for net working bet ween t he packet and circuit swit ched services.

GPRS specifies signaling t hat can enable support for various t raffic st ream s wit h different charact erist ics such as const ant bit rat e or variable bit rat e, connect ion or connect ionless, point t o- point or point - t o- m ult ipoint , and bi- direct ional or unidirect ional st ream s. These t raffic st ream s can be est ablished on a dem and, reserved, or perm anent basis. The st andards define several subscriber profiles t hat include inform at ion and param et ers based on t he cont ract ual service agreem ent bet ween t he user and t he carrier. The profile includes subscriber services and subscriber QoS profiles ( e.g., service priorit y, reliabilit y, delay, and t hroughput ) . To provide end- t o- end QoS, we need support from several com ponent s of t he m obile syst em . The GPRS prot ocol archit ect ure ( cont rol plane) is shown in Figure 10.24. We will focus on t wo QoS aspect s: t he GPRS bearer service QoS profile ( Sect ion 10.7.1) and radio bearer service QoS m echanism s ( Sect ion 10.7.2) .

Figu r e 1 0 .2 4 . GPRS Pr ot ocol Ar ch it e ct u r e ( Con t r ol Pla n e )

10.7.1 QoS Profile in GPRS Bearer Service I n order t o provide QoS t o an applicat ion, t he net work resources on each com ponent of t he m obile syst em ( i.e., m obile st at ion, Serving GPRS Support Node [ SGSN] , Gat eway GPRS Support Node [ GGSN] ) need t o be reserved. Resource reservat ion is carried out by t he signaling m echanism t hat uses QoS profiles. A QoS profile is t he set of QoS param et ers t hat describes t he applicat ion's charact erist ics and QoS requirem ent s. A QoS profile consist s of four param et ers: service precedence ( priorit y) param et er, reliabilit y param et er, delay param et ers, and t hroughput param et er. These param et ers are described below.

10.7.1.1 Service Precedence or Priority Parameter This includes t he priorit y of support ing a specific t raffic st ream com pared t o ot her t raffic st ream s. Higher priorit y st ream s will be served before lower priorit y st ream s. I n case t he net work is congest ed, lower priorit y packet s will be discarded before t he higher priorit y packet s. Three levels ( classes) of priorit ies are applied ( see Table 10.1) .

Ta ble 1 0 .1 . Se r vice Pr e ce de n ce Cla sse s Pr e ce de n ce

Pr e ce de n ce N a m e

1

High Priorit y

2

Norm al Priorit y

3

Low Priorit y

10.7.1.2 Reliability Parameter This indicat es t he t ransm ission charact erist ics required by t he applicat ion in t erm s of loss probabilit y, duplicat ion of packet s, m is- sequencing of packet s, or corrupt ion of packet s. GPRS defines t hree reliabilit y classes t hat are described in Table 10.2.

Ta ble 1 0 .2 . Re lia bilit y Cla sse s

Re lia bilit y Cla ss

Lost Pa ck e t Pr oba bilit y

D u plica t e Pa ck e t Pr oba bilit y

Ou t of Se qu e n ce Pa ck e t Pr oba bilit y

1

10 - 9

10 - 9

10 - 9

10 - 9

2

10 - 4

10 - 5

10 - 5

10 - 6

3

10 - 2

10 - 5

10 - 5

10 - 2

Cor r u pt Pa ck e t Pr oba bilit y

10.7.1.3 Delay Parameter This param et er defines t he m axim um values for t he m ean packet delay and t he 95 percent ile delay t hat t he GPRS net work can provide. This delay can consist of t he t ransm ission delay from t he user t o t he net work ( uplink) , or from t he net work t o t he user ( downlink) , and t he delay in t he GPRS backbone. This delay does not include ext ernal net works or user higher layers. GPRS defines five classes of delay values described in Table 10.3.

Ta ble 1 0 .3 . D e la y Cla sse s D e la y ( m a x im u m va lu e s) D e la y Cla ss

Pa ck e t Size : 1 2 8 oct e t s

Pa ck e t Size : 1 0 2 4 oct e t s

M e a n Tr a n sfe r D e la y ( se c)

9 5 Pe r ce n t ile D e la y ( se c)

M e a n Tr a n sfe r D e la y ( se c)

9 5 Pe r ce n t ile D e la y ( se c)

1. ( Predict ive)

< 0.5

< 1.5

< 2

< 7

2. ( Predict ive)

< 5

< 25

< 15

< 75

3. ( Predict ive)

< 50

< 250

< 75

< 375

4. ( Best Effort )

Unspecified

10.7.1.4 Throughput Parameter The t hroughput param et ers include t he m axim um and m ean bit rat es. These t erm s can be negot iat ed bet ween t he user and t he net work before and during t he com m unicat ion session.

10.7.1.5 PDP Context Procedures Each applicat ion's QoS requirem ent s are m apped t o a QoS profile. The m apping process is not defined in t he st andard. The QoS profile is included in t he Packet Dat a Prot ocol ( PDP) cont ext . QoS profiles negot iat ion is m anaged by PDP cont ext procedures ( i.e., act ivat ion, m odificat ion,

and deact ivat ion) . PDP cont ext procedures are carried out by t he Session Managem ent ( SM) prot ocol layer bet ween MS and SGSN and by t he GPRS Tunneling Prot ocol ( GTP) layer bet ween SGSN and GGSN as shown in Figure 10.25. Generally, PDP cont ext procedures are used in PDP address ( i.e., I P address) assignm ent and host configurat ion t o a m obile st at ion in addit ion t o QoS profile signaling. Each applicat ion on a m obile st at ion perform s t he PDP cont ext procedures separat ely.

Figu r e 1 0 .2 5 . PD P Con t e x t Pr oce du r e s

Figure 10.26 shows t he PDP cont ext act ivat ion procedure init iat ed by t he MS. First , t he MS sends SGSN an " Act ivat e PDP Cont ext Request " m essage t hat cont ains several inform at ion elem ent s as well as t he desired QoS profile of an applicat ion. The SGSN decides t o accept or rej ect t he request based on t he available resources at t he SGSN and subscribed QoS profile. I f t he SGSN accept s t he request , it will send GGSN a " Creat e PDP Cont ext Request " m essage cont aining t he QoS profile. Aft er receiving t he " Creat e PDP Cont ext Request " m essage, t he GGSN decides t o accept or rej ect t he request based on t he available resources at t he GGSN and t he m obile st at ion's subscript ion inform at ion. I f t he GGSN accept s t he request , GGSN sends t he SGSN a " Creat e PDP Cont ext Response" m essage including t he negot iat ed QoS profile. Finally, SGSN includes t he negot iat ed QoS profile in an " Act ivat e PDP Cont ext Accept " m essage and sends it t o t he MS.

Figu r e 1 0 .2 6 . PD P Con t e x t Act iva t ion Pr oce du r e I n it ia t e d by M S

There are cases where packet s originat ed from t he ext ernal net work are received by t he GGSN and are dest ined t o a m obile st at ion. I n t hese cases, t he GGSN will init iat e t he PDP cont ext act ivat ion procedure shown in Figure 10.27. I n St ep 2 t he GGSN acquires t he rout ing inform at ion t o t he MS from t he Hom e Locat ion Regist er ( HLR) . Aft er receiving t his rout ing inform at ion from t he HLR, in St ep 3 t he GGSN sends a " PDU Not ificat ion Request " m essage t o t he SGSN and t hen t he SGSN replies back wit h a " PDU Not ificat ion Response" t o confirm t hat t he SGSN will init iat e a " Request PDP Cont ext Act ivat ion" m essage. The SGSN sends t he MS t he " Request PDP Cont ext Act ivat ion" m essage. I n St ep 5, t he MS cont inues wit h a sim ilar process as in PDP cont ext act ivat ion init iat ed by t he MS case ( shown in Figure 10.26) .

Figu r e 1 0 .2 7 . PD P Con t e x t Act iva t ion Pr oce du r e I n it ia t e d by GGSN

10.7.2 QoS Mechanisms in the Radio Bearer Service

I n t his subsect ion we will focus on t he radio int erface QoS m echanism s t hat enable quant it at ive QoS services: classificat ion and packet scheduling. Figure 10.28 shows a sim plified diagram of t hese QoS m echanism s.

Figu r e 1 0 .2 8 . Sim plifie d GPRS QoS Ar ch it e ct u r e on t h e Ra dio I n t e r fa ce

First , an applicat ion perform s t he QoS negot iat ion wit h t he SGSN as described in t he previous subsect ion. When dat a packet s arrive from t he upper layer t o t he RLC/ MAC, t he TBF est ablishm ent process will be act ivat ed. During t he TBF est ablishm ent process, t he QoS profile will be m apped t o RLC/ MAC param et ers ( i.e., radio priorit y) included in t he PACKET CHANNEL REQUEST. The m apping process is not defined in t he st andard. Aft er t he TBF is est ablished, t he TBF QoS param et ers are used at t he MS t o classify t he packet s and at t he BSS t o m ake packet scheduling decisions. I n t he rem aining part of t his sect ion we provide m ore det ails on t he classificat ion and packet scheduling m echanism s.

10.7.2.1 Classification All packet s t hat belong t o a specific TBF are t agged wit h t he corresponding TFI . These t ags along wit h t he direct ion ( i.e., uplink, downlink) enable packet classificat ion. Therefore, GPRS uses per- flow classificat ion which support s per- flow QoS services. Not ice t hat due t o t he t em porary nat ure of t he TBF, an applicat ion m ay have t he TBF est ablished and t erm inat ed several t im es during t he course of t he service. Therefore, TFI will be different for each TBF connect ion.

10.7.2.2 Packet Scheduling As described in t he Sect ion 10.5.2, t he channel access is based on TDMA. The t im e slot assignm ent is based on t he request - grant process. The current bandwidt h dem and in t erm s of t he num ber of RLC/ MAC blocks wait ing in queue is included in t he PACKET CHANNEL REQUEST m essage sent during t he TBF est ablishm ent process. Packet scheduling uses bandwidt h dem and inform at ion as well as TBF QoS param et ers t o decide when packet s t hat belong t o a specific TBF are allowed t o t ransm it . Therefore, t he packet scheduling algorit hm can dynam ically allocat e t he bandwidt h based on an applicat ion's bandwidt h dem and and QoS requirem ent s. The dynam ic bandwidt h allocat ion decisions perform ed at t he BSS are com m unicat ed t o t he m obile st at ions using t he USF process ( see Sect ion 10.5.2) . I t is im port ant t o m ent ion t hat t he packet scheduling algorit hm s are not defined by t he st andard. These algorit hm s should be defined and im plem ent ed by t he GPRS net work designers and carriers. I n sum m ary, t he GPRS syst em com bined wit h proper packet classificat ion and packet scheduling algorit hm s can enable per- flow quant it at ive QoS services. < Day Day Up >

< Day Day Up >

Chapter 11. UMTS Sect ion 11.1. I nt roduct ion Sect ion 11.2. UMTS Archit ect ure Sect ion 11.3. Physical Layer Sect ion 11.4. Media Access Cont rol ( MAC) Sect ion 11.5. Dat a Link Layer Prot ocols ( RLC, PDCP, and BMC) Sect ion 11.6. Radio Resource Cont rol ( RRC) Sect ion 11.7. QoS Support < Day Day Up >

< Day Day Up >

11.1 Introduction Universal Mobile Telecom m unicat ions Syst em ( UMTS) is a 3G t echnology t hat provides highspeed connect ion of up t o 2 Mbps t o support a wide variet y of services, including m ult im edia services wit h different QoS requirem ent s. As shown in Figure 11.1, UMTS evolved from 2G GSM t echnology and 2.5G GPRS. The effect ive bandwidt h provided is up t o 144 kbps for vehicular users, 384 kbps for m obile users, and up t o 2 Mbps for st at ic users. The recent UMTS Rel- 5 added t o previous releases ( Rel- 99 and Rel- 4) t he High- Speed Download Packet Access ( HSDPA) t echnology. HSDPA adds an opt ion of higher speeds of up t o 10 Mbps required for m ult im edia applicat ions.

Figu r e 1 1 .1 . UM TS Evolu t ion

UMTS, which is also referred t o as WCDMA ( Wideband CDMA or Wideband Code Division Mult iple Access) , is part of I MT- 2000 ( I nt ernat ional Mobile Telecom m unicat ion–2000) effort , which is designed t o provide high- speed com m unicat ion wit h high- qualit y m ult im edia services and global roam ing support . The I MT- 2000 effort is led by t he I TU ( I nt ernat ional Telecom m unicat ions Union) , form erly t he CCI TT ( Consult at ive Com m it t ee for I nt ernat ional Telephony and Telegraphy) , is an int ernat ional organizat ion t hat set s com m unicat ion st andards and consist s of m ore t han 150 m em ber count ries. The I MT- 2000 effort st rives t o creat e a fam ily of com pat ible st andards t hat can be used worldwide for all m obile applicat ions. This fam ily is expect ed t o support bot h circuit and packet swit ched applicat ions and provides QoS support for m ult im edia t raffic. This effort considered several proposals for bot h t errest rial and sat ellit e com m unicat ion, from which WCDMA and CDMA 2000 were select ed for t errest rial com m unicat ion. UMTS has been int roduced and evolved in several releases. St art ing wit h Release 99 UMTS int egrat ed GSM and GPRS such t hat previous invest m ent s in t hose net works can be preserved. UMTS Terrest rial Radio Access Net work ( UTRAN) is UMTS's radio access net work t hat provides t he com m unicat ion pat h bet ween t he m obile st at ion and t he 3G core net work t hrough t he I u

int erface. UMTS shares t he sam e core net work as GPRS except t he UTRAN radio int erface, which is newly designed ( see Figure 11.2) . Therefore, GPRS and UMTS are fully harm onized.

Figu r e 1 1 .2 . GPRS a n d UM TS

UTRAN support s t wo m odes of operat ion: 1) Frequency Division Duplex ( FDD) , where uplink t ransm issions ( from t he m obile st at ions t o UTRAN) and downlink t ransm issions ( from UTRAN t o t he m obile st at ions) operat e sim ult aneously using t wo separat e frequency bands; and 2) Tim e Division Duplex ( TDD) , where uplink and downlink t ransm issions operat e int erchangeably in different t im e periods on t he sam e frequency band. I n UTRAN FDD m ode, UTRAN deploys Wideband Code Division Mult iple Access ( WCDMA) , whereas in UTRAN TDD m ode, UTRAN deploys Tim e Division–Code Division Mult iple Access ( TDCDMA) . Figure 11.3 shows UMTS coverage scenarios. FDD m ode, which is geared t oward wide area coverage ( i.e., public m acro and m icro cell) , support s high user m obilit y and dat a rat es of up t o 384 kbps. TDD, which is geared t oward sm aller geographical areas ( i.e., public m icro and pico cells) , support s dat a rat es of up t o 2 Mbps and slow user m obilit y ( i.e., cordless phones) . I n order t o operat e in bot h wide and local areas, a m obile st at ion or user equipm ent ( UE) should support bot h TDD and FDD m odes.

Figu r e 1 1 .3 . UM TS Ge ogr a ph ica l Cove r a ge

I n t his book we focus on UMTS Rel- 5, which is present ed in t he following docum ent s: ●

●

●

●

●

●

●

●

●

●

TS 23.002 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Services and Syst em s Aspect s; Net work Archit ect ure ( Release 5) TS 25.301 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Radio I nt erface Prot ocol Archit ect ure ( Release 5) TS 25.401 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; UTRAN Overall Descript ion ( Release 5) TS 25.302 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Services provided by t he physical layer ( Release 5) TS 25.211 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Physical channels and m apping of t ransport channels ont o physical channels ( FDD) ( Release 5) TS 25.321 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; MAC Prot ocol Specificat ion ( Release 5) TS 25.322 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Radio Link Cont rol ( RLC) Prot ocol Specificat ion ( Release 5) TS 25.323 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Packet Dat a Convergence Prot ocol ( PDCP) Specificat ion ( Release 5) TS 25.331 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Radio Resource Cont rol ( RRC) ; Prot ocol Specificat ion ( Release 5) TS 25.308 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; High- Speed Downlink Packet Access ( HSDPA) ; Overall Descript ion; St age 2 ( Release 5) < Day Day Up >

< Day Day Up >

11.2 UMTS Architecture UMTS Rel- 5 provides t he ult im at e evolut ion of voice and dat a convergence. As shown in Figure 11.4, sim ilar t o GPRS archit ect ure, UMTS consist s of t hree m ain ent it ies: Mobile St at ion ( MS) , UTRAN, and Core Net work ( CN) . A m obile st at ion or user equipm ent ( UE) com m unicat es wit h Node B which cont rols t he radio channel wit hin it s coverage area or cell. Mult iple Node Bs are cont rolled by t he Radio Net work Cont roller ( RNC) . UMTS's Node B is equivalent t o GPRS's Base Transceiver St at ion ( BTS) , while RNC is equivalent t o GPRS's Base St at ion Cont roller ( BSC) . RNC and associat ed Node B form t he radio net work subsyst em ( RNS) . The RNCs are int erconnect ed t o each ot her t hrough t he I ur int erface. RNC connect s t o t he core net work t hrough I u int erface which support s bot h voice and dat a services.

Figu r e 1 1 .4 . UM TS Ar ch it e ct u r e

UTRAN's prot ocol st ack present ed in Figure 11.5 consist s of t he Physical Layer ( Layer 1) , Dat a

Link layer ( Layer 2) , and Net work layer ( Layer 3) . The Physical Layer ( PHY) includes t he radio. The Dat a Link includes t he Media Access Cont rol ( MAC) , t he Radio Link Cont rol ( RLC) , t he Packet Dat a Convergence Prot ocol ( PDCP) , and t he Broadcast / Mult icast Cont rol ( BMC) . The Net work Layer includes t he Radio Resource Cont rol ( RRC) . The RRC and RLC are divided int o Cont rol and User planes. PDCP and BMC exist in t he User plane only. The PDCP and RLC in Rel5 are unchanged from t he Rel- 99 and Rel- 4 archit ect ure, whereas t he MAC is changed t o include a new m odule t hat support s HSDPA. Layers 1, 2, and 3 are part of UMTS's Access St rat um ( AS) . The higher layer t hat is part of UMTS's Non- Access St rat um ( NAS) includes Mobilit y Managem ent ( MM) , Call Cont rol ( CC) , and Session Managem ent ( SM) .

Figu r e 1 1 .5 . UTRAN 's Pr ot ocol St a ck

The t ransport channels which indicat e how t he dat a are t ransferred over t he radio int erface are defined by t he service access point ( SAP) bet ween t he MAC and PHY layers. The logical channels which indicat e t he cont ent s or t ypes of dat a t ransferred are defined by t he SAP bet ween t he RLC and MAC layers. Therefore, one of t he MAC's funct ions is t o m ap t he logical channels t o t he t ransport channels. The services provided by layer 2 are considered t o be t he radio bearer. The Physical Layer, which is described in det ail in t he next sect ion, provides t ransport services

t o t he MAC t hrough t he t ransport channels. The physical layer t ransport channels are m apped t o t he physical channels by defining a num ber of radio m echanism s ( i.e., m odulat ion, channel coding, radio m at ching, m ult iplexing, int erleaving) . The MAC layer cont ains several MAC ent it ies ( i.e., MAC- d, MAC- c/ sh) which will be described in Sect ion 11.4. The Radio Link Cont rol ( RLC) , t he Packet Dat a Convergence Prot ocol ( PDCP) , t he Broadcast / Mult icast Cont rol ( BMC) , and t he Radio Resource Cont rol ( RRC) are int roduced in Sect ion 11.5. < Day Day Up >

< Day Day Up >

11.3 Physical Layer The PHY provides dat a t ransfer services t o t he MAC and higher layers. The PHY is responsible for 1) various handover funct ions, 2) error det ect ion and report t o higher layers, 3) m ult iplexing of t ransport channels, 4) m apping of t ransport channels t o physical channels, 5) power cont rol, 6) synchronizat ion in TDD m ode, and 7) ot her radio responsibilit ies associat ed wit h t ransm it t ing and receiving signals over t he wireless m edia. The PHY's UMTS Terrest rial Radio Access Net work ( UTRAN) support s t he following t wo m odes: FDD ( Frequency Division Duplex) m ode and TDD ( Tim e Division Duplex) m ode. I n t he FDD m ode, t he PHY radio works in t wo frequency bands ( each band has 5 MHz channel bandwidt h) , one band for uplink t ransm ission and t he ot her for downlink t ransm ission. I n TDD m ode, t he sam e frequency band ( wit h a 5 MHz channel bandwidt h) is used for bot h uplink and downlink t ransm issions. The uplink and downlink operat e reciprocally in different t im e periods. Figures 11.6 and 11.7 illust rat e t he 3GPP ( 3G Part nership Proj ect ) frequency spect rum for t he FDD and TDD m odes, respect ively. However, som e count ries m ay define t heir own frequency bands for 3G.

Figu r e 1 1 .6 . UM TS FD D Fr e qu e n cy Ba n ds

Figu r e 1 1 .7 . UM TS TD D Fr e qu e n cy Ba n ds

The FDD m ode allows sym m et ric dat a t ransm issions on t he uplink and downlink. On t he ot her hand, t he TDD m ode has m ore flexibilit y in assigning asym m et ric uplink and downlink t ransm issions. Therefore, TDD is suit able t o support asym m et ric services, for exam ple where m ost of t he t raffic is on t he downlink ( from t he backbone t o t he m obile users) . However, since TDD requires t im e synchronizat ion, it cannot be deployed in wide coverage areas ( i.e., m ega cell) where t he large variat ions in propagat ion delays do not allow t ight synchronizat ion bet ween t he nodes. For channel access, t he PHY deploys Wideband Direct Sequence Code Division Mult iple Access ( WCDMA) , in which t he inform at ion bit s spread over 5 MHz channel bandwidt h using a num ber of chip signals. The 3.84 Mcps chip rat e is used in bot h FDD and TDD m odes. I n addit ion, TDD has an opt ion of 1.28 Mcps chip rat e, which spreads t he inform at ion bit s over 1.6 MHz channel bandwidt h. The channel access of t he lat t er case is denot ed as narrowband CDMA. The key concept of code division m ult iple access is t hat m ult iple users can t ransm it dat a sim ult aneously on t he sam e frequency band using different codes. Figure 11.8 illust rat es t he channel access of t he FDD and TDD m odes.

Figu r e 1 1 .8 . Ch a n n e l Acce ss

The code division m ult iple access involves t wo t ypes of codes: scram bling code and channelizat ion code. Figure 11.9 illust rat es t he sim plified diagram of t he coding process. The dat a from a physical channel are coded wit h channelizat ion code and scram bling code and t hen

t ransm it t ed. At t he receiver end, t he sam e codes ( bot h channelizat ion and scram bling) are used t o decode t he signal.

Figu r e 1 1 .9 . Ch a n n e liza t ion a n d Scr a m blin g Pr oce ss

The scram bling code is based on t he com plex- valued Gold Code som et im es called Pseudo Noise ( PN) code. The scram bling code is used t o separat e bet ween t ransm issions from different nodes. For exam ple, when m ult iple t ransm issions arrive sim ult aneously on t he uplink, Node B can ret rieve t he dat a signal from a specific m obile st at ion by applying t he sam e scram bling code of t he m obile st at ion t o t he incom ing signal. On t he downlink, where only Node Bs t ransm it , t he scram bling code can also be applied t o separat e t he dat a signals from different cells. For t he uplink, t here is a large num ber of scram bling codes ( i.e., in t he order of several m illions) t o be assigned t o t he m obile st at ions. For t he downlink, t he st andard lim it s t he num ber of codes t o only 512. One of t he reasons for t his lim it ed num ber of codes is t o reduce t he t im e it t akes a new st at ion t o j oin a cell. When a new st at ion j oins a cell, t he st at ion has no knowledge of what scram bling code is used by Node B. The st at ion has t o scan t he channel t o find out t his code. Obviously, searching only 512 codes, as opposed t o m illions of codes, expedit es t his search process. The channelizat ion code is based on Ort hogonal Variable Spreading Fact or ( OVSF) codes, som et im es called Walsh codes. The channelizat ion code is used t o separat e t he sources of different physical channels on a m obile st at ion. The physical channels will be described in a lat er sect ion. Different physical channels on a m obile st at ion deploy different OVSF codes. I t is possible t o have t he sam e OVSF code assigned t o physical channels on different m obile st at ions. An OVSF code t hat cont ains a sequence of chips spreads t he dat a signal ( in bit s) t o a higher bandwidt h dat a signal ( in chips) . OVSF codes are organized in a t ree st ruct ure as shown in Figure 11.10.

Figu r e 1 1 .1 0 . OVSF Code Tr e e

The OVSF code t ree cont ains t he hierarchical st ruct ure of codes where each level is referred t o t he spread fact or ( SF) . For exam ple, SF = 4 cont ains four OVSF codes wit h t he lengt h of four chips each. Codes wit hin t he sam e spread fact or are ort hogonal t o each ot her. The code assignm ent for t he uplink is m anaged by each m obile st at ion, and by t he RNC on t he downlink. There are som e rest rict ions for t he code assignm ent . A code can be assigned t o a physical channel if t here is no ot her physical channel t hat is allocat ed t his code or codes in t he underlying branch. For exam ple ( see Figure 11.10) , code C2,1 can be assigned if code C2,1 and codes in t he underlying branch ( i.e., C4,1 , C4,2 , C8,1 , C8,2 , C8,3 , C8,4 and so on) are not assigned t o any physical channel. Furt herm ore, sm aller spread fact or codes on t he pat h t o t he root cannot be used ( i.e., C1,1 ) . The level of signal spreading is det erm ined by t he spread fact or. For exam ple, a spread fact or equal t o 256 allows 256 OVSF codes ( 256 chip lengt h each code) . The higher t he spread fact or is, t he lower t he sym bol dat a rat e ( or bit rat e) but t he higher t he num ber of physical channels t hat can t ransm it sim ult aneously. The possible spread fact or num bers are 1, 2, 4, 8, 16, 32, and so on. I n FDD m ode, t he spread fact ors are from 256 t o 4 on t he uplink and from 512 t o 4 on t he downlink. Wit h 3.84 Mcps chip rat e, in FDD, t he inform at ion rat e varies from 15 ksym bols/ s ( 3.84 M/ 256) t o 960 ksym bols/ s ( 3.84 M/ 4) on t he uplink, and from 7.5 ksym bols/ s ( 3.84 M/ 512) t o 960 ksym bols/ s on t he downlink. I n TDD m ode, t he spread fact ors range from 16 t o 1 on bot h t he uplink and downlink. I n TDD, t he inform at ion rat e varies from 240 ksym bols/ s ( 3.84M/ 16) t o 3.84 Msym bols/ s. The assigned dat a rat e or bandwidt h t o a connect ion is cont rolled by t he spread fact or. The

spread fact or assignm ent can be fixed or dynam ically adj ust ed based on t he bandwidt h dem and of t he connect ion. The lower t he spread fact or is, t he higher t he achievable dat a rat e of t he connect ion. Anot her approach t o provide higher dat a rat es is t o assign m ult iple physical channels t o a connect ion. An im port ant issue we need t o consider in WCDMA is t he power cont rol ( especially on t he uplink) . Aft er t he incom ing signal is appropriat ely decoded ( i.e., scram bling and channelizat ion decoding) , t he desired signal is em inent from ot her signals ( which are considered as noise int erference) . I f t he power densit y of t he undesired signals ( ot her channels) is t oo high, t he decoded signal will cont ain a high level of noise int erference and event ually com prom ise t he desired signal det ect ion. Therefore, wit hout power cont rol, an overpowered st at ion could int erfere or block t he signals from ot her st at ions. Figure 11.11 illust rat es power cont rol in WCDMA.

Figu r e 1 1 .1 1 . Pow e r Con t r ol in UM TS

The st andard defines t wo t ypes of power cont rol: open- loop power cont rol and fast closed- loop power cont rol. I n t he open- loop power cont rol m echanism , a m obile st at ion est im at es t he signal st rengt h from t he downlink beacon signal and set s it s uplink st rengt h power. This power cont rol is inaccurat e and oft en used in init ial power set t ing. I n t he fast closed- loop power cont rol m echanism , Node B m easures t he signal- t o- int erference rat io ( SI R) from t he m obile st at ion. Then, Node B report s t his SI R m easurem ent t o t he RNC, which in t urn det erm ines t he appropriat e value of t he m obile st at ion's signal power. The RNC sends t he Transm it Power Cont rol ( TPC) com m and ( included in t he physical cont rol channel) t o t he m obile st at ion. The m obile st at ion readj ust s t he t ransm ission power according t o t he received TPC com m and. UTRAN deploys t he QPSK m odulat ion schem e. 8- PSK is used for 1.28 Mcps TDD opt ion, whereas 16- QAM is used in HS- DSCH.

11.3.1 Transport and Physical Channels One of t he funct ions of t he physical layer is t o perform t he m apping process bet ween t he t ransport and physical channels. As shown in Figure 11.12, t he MAC packet s on a t ransport channel are called t ransport blocks. A num ber of t ransport blocks t ransm it t ed on t he sam e

t ransport channel are called a Transport Block Set . A Transport Block Set size is defined as t he num ber of bit s in a Transport Block Set . As shown in Figure 11.13, t he int erarrival t im e bet ween consecut ive Transport Block Set s is defined as t he Transm ission Tim e I nt erval ( TTI ) .

Figu r e 1 1 .1 2 . Tr a n spor t Block

Figu r e 1 1 .1 3 . Tr a n spor t Tim e I n t e r va l

The Transport Block Size, t he Transport Block Set Size, and t he Transm ission Tim e I nt erval are included in t he Transport Form at . The Transport Form at cont ains t wo part s: t he dynam ic and sem i- st at ic part s. The dynam ic part cont ains t he Transport Block Size and t he Transport Block Set Size. The sem i- st at ic part cont ains t he Transm ission Tim e I nt erval, t he error prot ect ion schem e ( channel code, coding rat e, rat e m at ching) , and t he CRC size. There are a num ber of Transport Form at com binat ions called Transport Form at Com binat ion Set ( TFCS) . For exam ple,

a TFCS is shown in Tables 11.1 and 11.2 for dynam ic and sem i- st at ic part s, respect ively. The TFCS is assigned by t he Radio Resource Cont rol ( RRC) . The MAC select s an appropriat e Transport Form at Com binat ion ( TFC) from t he TFCS t o support t he connect ion's t raffic charact erist ics and dat a rat e. The MAC uses dynam ic at t ribut es for t he TFC and TTI , while t he PHY uses sem i- st at ic at t ribut es for t he TFC.

Ta ble 1 1 .1 . Tr a n spor t For m a t Com bin a t ion Se t D yn a m ic Pa r t Tr a n spor t Ch a n n e l 1

Tr a n spor t Ch a n n e l 2

Tr a n spor t Ch a n n e l 3

Tr a n spor t block size ( bit s)

Tr a n spor t block se t size ( bit s)

Tr a n spor t block size ( bit s)

Tr a n spor t block se t size ( bit s)

Tr a n spor t block size ( bit s)

Tr a n spor t block se t size ( bit s)

Com binat ion 1

20

20

100

200

120

240

Com binat ion 2

40

40

160

160

320

320

Com binat ion 3

320

1280

40

40

320

1280

Ta ble 1 1 .2 . Tr a n spor t For m a t Com bin a t ion Se t Se m i- St a t ic Pa r t Tr a n spor t Tim e I n t e r va l ( m s)

Type of Er r or Pr ot e ct ion Code

St a t ic Ra t e M a t ch in g Pa r a m e t e r

Transport Channel 1

10

Convolut ion code

1

Transport Channel 2

20

Turbo code

1

Transport Channel 3

30

Turbo code

2

Figure 11.14 shows t he physical layer m apping process. I n a cert ain period of t im e, t ransport blocks arrive from t he MAC along wit h Transport Form at I ndicat or ( TFI ) which indicat es t he t ransport form at of t he t ransport blocks. The t ransport blocks from different t ransport channels are m ult iplexed, channel coded ( i.e., convolut ion code, t urbo code) , split , and passed t o t he physical dat a channels. The TFI s are also m ult iplexed ont o t he Transport Form at Com binat ion I ndicat or ( TFCI ) and t hen passed t o a physical cont rol channel. There can be one or m ult iple physical dat a channels but t here is only one physical cont rol channel. At t he receiver end, aft er descram bling and channelizat ion decoding, t he receiver passes t he incom ing packet s t o t he decoding, dem ult iplexing, and channel j oining m odule which uses t he inform at ion wit hin t he TFCI . The t ransport blocks are rout ed t o t he appropriat ed t ransport channels t o t he higher layer.

Figu r e 1 1 .1 4 . M a ppin g Pr oce ss a t t h e Ph ysica l La ye r

11.3.1.1 Transport channels The t ransport channels ( see Figure 11.15) t hat t ransm it t he dat a t hrough t he radio int erface can be cat egorized as eit her dedicat ed t ransport channels or com m on t ransport channels. Dedicat ed t ransport channels are ident ified by t he physical channel—t hat is, 1) code and frequency in FDD m ode, or 2) code and t im e slot in TDD m ode. Com m on t ransport channels are ident ified ( if required) by an in- band ident ificat ion.

Figu r e 1 1 .1 5 . Tr a n spor t Ch a n n e ls

Since net work resources are reserved for t he dedicat ed t ransport channels, t hese channels are suit able for real- t im e applicat ions. Dedicat ed Channel ( DCH) is a channel dedicat ed t o one m obile st at ion and is used in eit her uplink or downlink direct ion. DCH m ay cont ain user or cont rol dat a from t he upper layer. Com m on t ransport channels include: ●

●

●

●

●

●

●

●

Broadcast Channel ( BCH) is a downlink channel used t o broadcast syst em inform at ion t o t he ent ire cell such as t he available access code and access t im e slot s used for t he random access m echanism . BCH's t ransm ission power should be sufficient t o reach all m obile st at ions wit hin a cell. Paging Channel ( PCH) is a downlink channel used t o broadcast cont rol inform at ion t o t he ent ire cell in order for t he m obile st at ion t o execut e it s sleep m ode procedures. Forward Access Channel ( FACH) is a com m on downlink channel used t o t ransm it relat ively sm all am ount s of dat a ( i.e., SMS m essage) or cont rol m essages. FACH m ay cont ain m essages dest ined t o several m obile st at ions. Therefore, in- band ident ificat ion is required. FACH provides low bit rat e and does not perform fast closed- loop power cont rol. Downlink Shared Channel ( DSCH) is a downlink channel shared by several m obile st at ions. I t is used for t ransm ission of dedicat ed user or cont rol dat a. DSCH provides variable bit rat e and is required t o perform fast closed- loop power cont rol. High- Speed Downlink Shared Channel ( HS- DSCH) is a downlink channel shared bet ween m obile st at ions by allocat ion of individual codes, from a com m on pool of codes assigned for t he channel. I t is sim ilar t o DSCH but provides higher dat a rat es. Random Access Channel ( RACH) is a cont ent ion based uplink channel used for t ransm ission of relat ively sm all am ount s of dat a ( e.g., for init ial access or non- real- t im e cont rol or user dat a) . RACH does not perform fast closed- loop power cont rol. Due t o unpredict able access delays of t he cont ent ion based schem e, RACH is not suit able for real- t im e cont rol and user dat a. Com m on Packet Channel ( CPCH) is a cont ent ion based uplink channel shared by t he m obile st at ions used for t ransm ission of burst y dat a t raffic. CPCH provides variable bit rat e and is required t o perform fast closed- loop power cont rol. ( Not e: This channel is used in FDD m ode only.) Uplink Shared Channel ( USCH) is an uplink channel shared by several m obile st at ions. I t is used for dedicat ed cont rol or user dat a. ( Not e: This channel is used in TDD m ode only.)

I n sum m ary, t raffic can be t ransm it t ed t hrough: ●

Uplink - RACH: cont ent ion based, sm all am ount s of dat a, no power cont rol - CPCH ( FDD m ode only) : cont ent ion based, m ore dat a t han RACH, power cont rol required - USCH ( TDD m ode only) - DCH: cont ent ion free, dedicat ed bandwidt h, large am ount s of dat a, power cont rol required and QoS support

●

Downlink - FACH: sm all am ount s of dat a, shared wit h m ult iple users

- DSCH: m ore dat a t han FACH, shared wit h m ult iple users - DCH: dedicat ed bandwidt h t o a user and QoS support

11.3.1.2 Physical Channels The physical channels are defined in t he physical layer by a specific channel frequency, channelizat ion code, and t im e durat ion. The st andard defines t he st ruct ure of t he physical channel in radio fram es and t im e slot s. A radio fram e has t he durat ion of 38400 chips, which is equivalent t o 10 m s at 3.84 Mcps chip rat e. A radio fram e consist s of 15 t im e slot s and t he durat ion of each t im e slot corresponds t o 2560 chips. Each t im e slot is num bered from 0 t o 14 ( see Figure 11.16) . This fram e st ruct ure is applied t o bot h TDD and FDD m odes. A physical channel can be t ransm it t ed on one or m ult iple t im e slot s ( consecut ively or nonconsecut ively) . Mult iple physical channels wit h different channelizat ion codes can coexist in t he sam e t im e slot . Som e physical channels m ay have different st ruct ures.

Figu r e 1 1 .1 6 . Ph ysica l Ch a n n e l Fr a m e St r u ct u r e

Figure 11.17 shows a fram e st ruct ure of t he UTRAN TDD m ode wit h several swit ching point configurat ions.

Figu r e 1 1 .1 7 . UTRAN TD D Fr a m e St r u ct u r e

The st andard defines several physical channels t o support t he t ransport channels. Som e physical channels are used for Layer 1 cont rol exchange bet ween UE and UTRAN and are not visible t o t he upper layer. The FDD physical channels include: ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

Dedicat ed Physical Dat a Channel ( DPDCH) Dedicat ed Physical Cont rol Channel ( DPCCH) Physical Random Access Channel ( PRACH) Physical Com m on Packet Channel ( PCPCH) Com m on Pilot Channel ( CPI CH) Prim ary Com m on Cont rol Physical Channel ( P- CCPCH) Secondary Com m on Cont rol Physical Channel ( S- CCPCH) Synchronizat ion Channel ( SCH) Physical Downlink Shared Channel ( PDSCH) Acquisit ion I ndicat or Channel ( AI CH) Access Pream ble Acquisit ion I ndicat or Channel ( AP- AI CH) Paging I ndicat or Channel ( PI CH) CPCH St at us I ndicat or Channel ( CSI CH) Collision- Det ect ion/ Channel- Assignm ent I ndicat or Channel ( CD/ CA- I CH) High- Speed Physical Downlink Shared Channel ( HS- PDSCH) HS- DSCH- relat ed Shared Cont rol Channel ( HS- SCCH)

The TDD physical channels include: ● ● ● ● ● ● ● ●

Dedicat ed Physical Channel ( DPCH) Prim ary Com m on Cont rol Channel Physical Channel ( P- CCPCH) Secondary Com m on Cont rol Physical Channel ( S- CCPCH) Physical Random Access Channel ( PRACH) Physical Uplink Shared Channel ( PUSCH) Physical Downlink Shared Channel ( PDSCH) Paging I ndicat or Channel ( PI CH) Synchronizat ion Channel ( SCH)

Figure 11.18 shows t he m apping of t he t ransport channels t o t he physical channels.

Figu r e 1 1 .1 8 . Tr a n spor t Ch a n n e ls M a ppe d on t o Ph ysica l Ch a n n e ls

I n t his chapt er we cover t he following m echanism s: dedicat ed dat a t ransm issions and random access uplink dat a t ransm issions.

11.3.2 Dedicated Data Transmission DCH which support s dedicat ed dat a t ransm issions ( uplink and downlink) is m apped t o DPDCH and DPCCH as shown in Figure 11.19.

Figu r e 1 1 .1 9 . D CH M a ppin g Pr oce ss a n d Ph ysica l Ch a n n e l Fr a m e St r u ct u r e

11.3.2.1 Uplink DPCCH carries cont rol inform at ion generat ed at t he physical layer. A pilot bit is used for channel est im at ion in coherent det ect ion. Transport Form at Com binat ion I ndicat or ( TFCI ) not ifies t he receiver about t he form at com binat ion of t ransport channels m apped t o DPDCHs. Transm it Power Cont rol ( TPC) is used for fast close- loop power cont rol. DPDCH cont ains user dat a ( t ransport block) from t he upper layer. FBI cont ains feedback inform at ion. DPCCH uses a fixed channelizat ion code wit h SF = 256. Therefore, a DPCCH fram e cont ains 10 bit s ( 2560/ 256) . DPDCH carries t he user dat a ( t ransport blocks) from DCHs. There are up t o six DPDCHs in a t im e slot . I f only one DPDCH exist s, t he DPDCH uses a channelizat ion code wit h SF ranging from 256 t o 4. I f m ult iple DPDCHs exist , DPDCHs use a channelizat ion code wit h SF = 4. I t is flexible t o vary t he dat a rat e of dedicat ed dat a t ransm issions using varying channelizat ion code ( from fram e t o fram e) and a num ber of concurrent DPDCHs. On t he uplink, each t im e slot cont ains one or m ore DPDCHs ( up t o six) as well as a DPCCH.

11.3.2.2 Downlink DPCH, t he downlink dedicat ed physical channel, cont ains t he m ult iplexing of DPDCHs and DPCCHs. DPCH uses a channelizat ion code wit h SF ranging from 512 t o 4.

11.3.3 Random Access Uplink Data Transmission RACH and CPCH support random access dat a t ransm ission ( uplink only) . CPCH is used only in FDD m ode. RACH is m apped t o PRACH whereas CPCH is m apped t o PCPCH. Figures 11.20 and 11.21 show t he m apping process of RACH and CPCH, respect ively. Each m obile st at ion has only one RACH and one CPCH. I n RACH, t he cont rol part cont ains TFCI t hat provides t he decoding inform at ion used for t he dat a part of t he PRACH. I n CPCH, t he cont rol part also cont ains TPC ( for power cont rol) , pilot , and FBI in addit ion t o TFCI . Bot h cont rol and dat a part use different channelizat ion codes.

Figu r e 1 1 .2 0 . RACH M a ppin g Pr oce ss a n d PRACH Fr a m e St r u ct u r e

Figu r e 1 1 .2 1 . CPCH M a ppin g Pr oce ss a n d PCPCH Fr a m e St r u ct u r e

The size of t he pream ble part of RACH and CPCH is 4096 chips and consist s of 256 repet it ions of a 16- chip signat ure. There are a m axim um of 16 16- chip signat ures available. The pream ble part is coded wit h a scram bling code ( Gold Code) . PRACH's cont rol part deploys a channelizat ion code wit h SF = 256 while PRACH's dat a part deploys a channelizat ion code wit h SF ranging from 256 t o 32. PCPCH's cont rol part deploys a fixed channelizat ion code wit h SF = 256 while PCPCH's dat a part deploys a channelizat ion code wit h SF ranging from 256 t o 4.

11.3.3.1 Random Access Uplink Data Transmission Procedure (for RACH) The Random Access Uplink Dat a Transm ission is based on slot t ed- Aloha wit h fast acquisit ion indicat ion. The dat a are t ransm it t ed at t he beginning of an access slot . The access slot s are spaced 5120 chips apart . There are 15 access slot s spanning on t wo radio fram es ( 2 x 10 m s = 20 m s) as shown in Figure 11.22. The Acquisit ion I ndicat or Channel ( AI CH) is a physical channel used for acquisit ion indicat ion. AI CH is defined at t he base st at ion whereas PRACH is defined at t he m obile st at ion. AI CH and PRACH access slot s are not aligned.

Figu r e 1 1 .2 2 . RACH Acce ss Slot s

As shown in Figure 11.23, RACH procedure includes t he following st eps: ●

●

●

●

●

●

●

A m obile st at ion ret rieves t he available RACH subchannel, scram bling codes, and signat ures from BCH. The m obile st at ion select s a RACH subchannel random ly and also furt her select s a signat ure from t he pool of available signat ures. RACH pream ble t hat cont ains t he select ed signat ure is t ransm it t ed on a random ly select ed access slot . I f t he m obile st at ion does not receive t he AI CH pream ble from t he base st at ion aft er t he RACH pream ble t im e out , t he m obile st at ion will ret ransm it t he RACH pream ble wit h higher t ransm ission power on t he next available access slot . I f t he num ber of ret ransm issions reaches t he m axim um num ber of ret ransm issions, t he RACH procedure st ops and not ifies t he upper layer. I f t he m obile st at ion receives t he AI CH pream ble wit h NACK ( negat ive acknowledgm ent ) , m eaning t hat t he select ed signat ure is already used by anot her st at ion, t he RACH procedure st ops and not ifies t he upper layer. I f t he m obile st at ion receives t he AI CH pream ble wit h ACK ( posit ive acknowledgm ent ) , t he m obile st at ion t ransm it s RACH m essages wit h a durat ion of 10 or 20 m s. The backoff m echanism is cont rolled by t he upper layer.

Figu r e 1 1 .2 3 . RACH Pr oce du r e

11.3.3.2 Random Access Uplink Data Transmission Procedure (for CPCH)

Random Access Uplink Dat a Transm ission for CPCH is sim ilar t o t hat for RACH. The difference is t hat CPCH uses AP- AI CH and CD/ CA- I CH inst ead of AI CH. I n addit ion, CPCH includes a collision det ect ion m echanism ( CD) t o reduce t he probabilit y of collision. The access slot s are sim ilar t o t hose for RACH ( see Figure 11.24) . AP- AI CH and CD/ CA- I CH are defined at t he base st at ion whereas PCPCH is defined at t he m obile st at ion. AP- AI CH and CD/ CA- I CH access slot s align but do not align wit h PCPCH access slot s.

Figu r e 1 1 .2 4 . CPCH Acce ss Slot s

As shown in Figure 11.25 t he CPCH procedure includes t he following st eps: ●

● ●

●

The procedure is sim ilar t o RACH procedure up unt il t he m obile st at ion receives AP- AI CH pream ble wit h ACK. The m obile st at ion t ransm it s Collision Det ect ion ( CD) pream ble t o t he base st at ion. The base st at ion echoes back wit h CD indicat ion. This handshaking process reduces t he probabilit y of collision. The m obile st at ion t ransm it s a CPCH m essage.

Figu r e 1 1 .2 5 . CPCH Pr oce du r e

< Day Day Up >

< Day Day Up >

11.4 Media Access Control (MAC) The MAC provides t he m apping process bet ween logical and t ransport channels. Before we describe t he MAC archit ect ure, we int roduce t he logical channels.

11.4.1 Logical Channels Logical channels ( see Figure 11.26) are channels t hat relat e t o t he cont ent and what kind of dat a are t ransm it t ed t hrough t he radio int erface. The Logical Channels are cat egorized int o t wo t ypes based on t heir cont ent : Cont rol Channels, which cont ain cont rol inform at ion, and Traffic Channels, which cont ain user dat a.

Figu r e 1 1 .2 6 . Logica l Ch a n n e ls

The Cont rol Channels ( CCHs) include t he following: ●

●

●

●

●

Broadcast Cont rol Channel ( BCCH) is a downlink channel for broadcast ing syst em cont rol inform at ion. Paging Cont rol Channel ( PCCH) is a downlink channel for paging inform at ion. This channel is used when t he net work does not know t he locat ion cell of t he m obile st at ion, or when t he m obile st at ion is in t he cell connect ed st at e using sleep m ode procedures. Com m on Cont rol Channel ( CCCH) is a bidirect ional channel for t ransm it t ing cont rol inform at ion bet ween t he net work and t he m obile st at ion. This channel is com m only used by a m obile st at ion wit h no RRC connect ion wit h t he net work and by m obile st at ions using com m on t ransport channels when accessing a new cell aft er cell reselect ion. Dedicat ed Cont rol Channel ( DCCH) is a point - t o- point bidirect ional channel t hat t ransm it s dedicat ed cont rol inform at ion bet ween t he m obile st at ion and t he net work. This channel is est ablished t hrough t he RRC connect ion set up procedure. Shared Channel Cont rol Channel ( SHCCH) is a bidirect ional channel t hat t ransm it s cont rol inform at ion for uplink and downlink shared channels bet ween t he net work and t he m obile st at ions. ( Not e: This channel is used in TDD m ode only.)

The Traffic Channels ( TCHs) include t he following: ●

●

Dedicat ed Traffic Channel ( DTCH) is a point - t o- point channel dedicat ed t o a single m obile st at ion for t he t ransfer of user inform at ion. A DTCH can exist in bot h uplink and downlink. Com m on Traffic Channel ( CTCH) is a point - t o- m ult ipoint unidirect ional channel for t ransfer of dedicat ed user inform at ion for all or a group of specified m obile st at ions.

The logical channels can be m apped ont o t he t ransport channels in t he MAC layer as shown in Figure 11.27.

Figu r e 1 1 .2 7 . Logica l Ch a n n e ls M a ppe d on t o Tr a n spor t Ch a n n e ls

11.4.2 MAC Architecture Figure 11.28 present s t he m obile st at ion MAC archit ect ure ( Rel- 5) and Figure 11.29 present s UTRAN MAC archit ect ure. The MAC consist s of t he following m odules: MAC- d, MAC- c/ ch and MAC- b. I n addit ion, UMTS Release 5 int roduced a new MAC- hs m odule t hat support s HSDPA's high- speed m ode.

Figu r e 1 1 .2 8 . M obile St a t ion M AC Ar ch it e ct u r e

Figu r e 1 1 .2 9 . UTRAN Side M AC Ar ch it e ct u r e

The MAC execut es t he following funct ions: ● ●

Maps bet ween t he logical and t ransport channels. Select s t he appropriat e Transport Form at , defined in Sect ion 11.3, from t he Transport Form at Com binat ion Set assigned by t he RRC for each Transport Channel depending on t he inst ant aneous source rat e. The Transport Form at ( which includes t he Transport Block Size, t he Transport Block Set Size, and t he Transport Tim e I nt erval)

●

●

●

●

●

●

● ●

●

●

accom m odat es t he source bandwidt h dem and. Priorit izes bet ween t he dat a flows of a m obile st at ion. Priorit y considerat ions can be achieved by select ing a Transport Form at Com binat ion for which t he high- priorit y dat a are m apped int o a " high bit rat e" Transport Form at , while lower priorit y dat a are m apped int o a " low bit rat e" Transport Form at . Priorit izes bet ween m obile st at ions via dynam ic scheduling. The MAC handles priorit ies on t he com m on and shared t ransport channels. For dedicat ed t ransport channels, t he equivalent of t he dynam ic scheduling funct ion is included as part of t he RRC. I dent ifies m obile st at ions on com m on t ransport channels via in- band ident ificat ion ( i.e., UE I D) . Mult iplexes and dem ult iplexes upper layer packet s int o/ from t ransport blocks delivered t o/ from t he physical layer on com m on t ransport channels and dedicat ed t ransport channels. Measures t raffic volum e on t he logical channels and report s it t o t he RRC. The RRC m ay swit ch t ransport channels based on t hese m easurem ent s. Swit ches bet ween com m on and dedicat ed t ransport channels based on t he swit ching decision m ade by t he RRC. Encrypt s dat a for securit y purposes. Accesses RACH and CPCH based on Service Classes. The RACH resources ( access slot s and pream ble signat ures for FDD, t im eslot and channelizat ion code for TDD) and CPCH resources ( access slot s and pream ble signat ures for FDD only) m ay be divided bet ween different Access Service Classes ( ASCs) in order t o provide different priorit ies of RACH and CPCH. This MAC funct ion indicat es t he appropriat e backoff t o t he physical layer for t he RACH and CPCH associat ed wit h t he MAC packet t ransfer. Execut es HARQ ( Hybrid Aut om at ic Repeat Request ) funct ionalit y for HS- DSCH t ransm ission t o ensure recept ion of all packet s. This funct ion is carried by part of t he MAC referred t o as t he MAC- hs. Delivers and assem bles higher layer packet s on HS- DSCH.

The execut ion of t he aforem ent ioned funct ions is undert aken by t he following MAC ent it ies: MAC- b for broadcast channels, MAC- c/ sh for com m on channels, MAC- d for dedicat ed channels, and MAC- hs for t he high- speed downlink shared channel. These MAC ent it ies have different funct ionalit y at t he m obile st at ion and at t he UTRAN.

11.4.2.1 MAC-b MAC- b m aps t he BCCH logical channel t o t he BCH t ransport channel.

11.4.2.2 MAC-c/sh MAC- c/ sh handles t he following t ransport channels: paging channel ( PCH) , forward access channel ( FACH) , random access channel ( RACH) , uplink com m on packet channel ( CPCH) t hat exist s only in FDD m ode, downlink shared channel ( DSCH) , and uplink shared channel ( USCH) t hat exist s only in TDD m ode. The m apping of logical channels on t ransport channels depends on t he m ult iplexing t hat is configured by t he RRC. I n t he downlink, if t he logical channels of t he dedicat ed t ype are m apped t o t he com m on t ransport channels, MAC- d receives t he dat a from MAC- c/ sh or MAC- hs via t he illust rat ed connect ion bet ween t he funct ional ent it ies. I n t he uplink, if t he logical channels of t he dedicat ed t ype are m apped t o t he com m on t ransport channels, MAC- d subm it s t he dat a t o MAC- c/ sh via t he illust rat ed connect ion bet ween t he funct ional ent it ies.

The MAC cont rols t he t im ing of RACH and CPCH t ransm issions. The physical RACH resources m ay be divided bet ween eight different ASCs in order t o provide different priorit ies of RACH usage. When t he MAC det erm ines t hat t he packet has not been received based on t he ACK, t he MAC init iat es ret ransm issions and backoff t im ing procedures before t rying t o ret ransm it . The num ber of ret ransm issions and t he backoff t im ings are based on ASC considerat ions.

11.4.2.3 MAC-d MAC- d handles t he dedicat ed t ransport channel ( DCH) t o MAC- c/ sh and MAC- hs.

11.4.2.4 MAC-hs MAC- hs handles t he high- speed downlink shared channel ( HS- DSCH) . I t is a new ent it y int roduced in Rel- 5 t hat support s HSDPA ( High- Speed Downlink Packet Access) .

11.4.3 Examples of MAC Data Transmission I n t his sect ion we describe a num ber of exam ples of t he MAC dat a t ransm ission ( Figure 11.30) . The logical channel for user dat a is DTCH, which can be m apped t o a dedicat ed t ransport channel ( i.e., DCH) , or com m on t ransport channel ( i.e., RACH and CPCH for uplink, FACH for downlink) , or shared t ransport channel ( i.e., USCH for uplink, DSCH for downlink) . I f DTCH is m apped t o a dedicat ed t ransport channel, t he dat a t ravel t hrough t he MAC- d m odule, whereas if DTCH is m apped t o a com m on t ransport channel or shared t ransport channel, t he dat a t ravel t hrough t he MAC- d and MAC- c/ sh m odules. For HS- DSCH, t he dat a t ravel t hrough t he MAC- d and MAC- hs m odules. Figure 11.30A shows t he uplink dat a flow when DTCH is m apped t o RACH. The diagram present s only t he channels involved in t his flow. The dat a flow is present ed in bold lines. Figure 11.30B illust rat es t he packet form at at each layer.

Figu r e 1 1 .3 0 . Ex a m ple of M AC D a t a Tr a n sm ission

The MAC SDU from DTCH visit s t he following m odules sequent ially: 1. Transport Channel Type Swit ching m odule, which decides t o which t ransport channels t o m ap it . The m apping is based on t he t raffic charact erist ics and QoS requirem ent s ( RACH for t his exam ple) . 2. C/ T MUX m odule, which is used when m ult iple dedicat ed logical channels are m apped t o a t ransport channel. C/ T header is added t o t he SDU t o provide an ident ificat ion of which logical channel t he packet belongs t o. 3. MAC- c/ sh Add/ Read UE I D m odule, which appends t o t he packet t he UE I D header t hat includes t he m obile st at ion ident ificat ion. The UE Type header ( 2 bit s) provides inform at ion of what kind of UE I D is used ( i.e., C- RNTI 16 bit s or U- RNTI 32 bit s) . 4. TCTF MUX m odule, where t he packet is m ult iplexed wit h packet s from ot her com m on logical channels. TCTF header is also added t o t he packet , which includes t he ident ificat ion of t he logical channel t he packet belongs t o. 5. Scheduling/ Priorit y m odule, which provides priorit y services. 6. Transport Form at ( TF) Select ion m odule, which select s a t ransport form at from t he

t ransport form at com binat ion set . 7. ASC Select ion m odule, which perform s t he priorit y random access procedure. < Day Day Up >

< Day Day Up >

11.5 Data Link Layer Protocols (RLC, PDCP, and BMC) 11.5.1 Radio Link Control (RLC) The RLC t ransfers dat a from t he upper layers in t ransparent or non- t ransparent m ode. I n t ransparent m ode upper layer packet s are t ransferred wit hout adding any prot ocol inform at ion. I n non- t ransparent m ode t he RLC will segm ent and reassem ble packet s and m ay concat enat e and pad packet s t o adj ust t he packet s t o t he t ransport form at . I n addit ion, t he RLC can t ransfer dat a in eit her unacknowledged or acknowledged m ode. I n unacknowledged dat a t ransfer, t he RLC t ransm it s upper layer packet s wit hout guarant eeing delivery t o addressee. I n acknowledged dat a t ransfer, t he RLC guarant ees delivery t o t he addressee. When t he packet cannot be t ransm it t ed, t he sender is being not ified. The RLC uses sequence num bers t o ident ify in- sequence and out - of- sequence delivery t o allow resequencing of packet s. I n case errors are ident ified, t he RLC request s ret ransm issions from t he sender. The RLC m aint ains QoS as defined by t he upper layers by providing different levels of service by m eans such as 1) not ifying t he upper layer of errors t hat cannot be resolved by t he RLC and 2) adj ust ing t he m axim um num ber of ret ransm issions according t o QoS delay requirem ent s.

11.5.2 Packet Data Convergence Protocol (PDCP) PDCP execut es header com pression and decom pression of I P dat a st ream s ( e.g., TCP/ I P and Real Tim e Prot ocol ( RTP) / UDP/ I P headers) at t he t ransm it t ing and receiving ent it y. I t t ransfers user dat a from t he upper layers and forwards t he dat a t o t he RLC layer and vice versa and m aint ains PDCP sequence num bers in som e cases.

11.5.3 Broadcast/Multicast Control (BMC) BMC support s broadcast / m ult icast t ransm ission services in t he user plane for com m on user dat a in unacknowledged m ode. The BMC st ores broadcast m essages for scheduled t ransm ission and t ransm it s t hese m essages at t he scheduled t im e. I n t he m obile st at ion side t he BMC also evaluat es t he scheduled m essages t o indicat e t he scheduling param et ers t o t he RRC. The BMS also m onit ors t raffic volum e, calculat es required resources and request s from t he RRC t hese resources on CTCH and FACH. The BMS in t he m obile st at ion receives t hese broadcast m essages and delivers error- free m essages t o t he upper layers. < Day Day Up >

< Day Day Up >

11.6 Radio Resource Control (RRC) RRC execut es t he cont rol plane signaling of t he net work layer bet ween t he m obile st at ions and t he UTRAN net work. The RRC perform s t he following funct ions: ●

● ●

●

●

● ●

●

Est ablishes, re- est ablishes, m aint ains, and releases RRC connect ions bet ween t he m obile st at ion and UTRAN. The est ablishm ent of an RRC connect ion is init iat ed by a request from t he m obile st at ion higher layers. Mult iple links can be est ablished t o a m obile st at ion sim ult aneously. I n t his process, t he RRC perform s adm ission cont rol and select s param et ers describing t he radio link based on inform at ion from higher layers. The RRC allocat es t he radio resource on t he uplink and downlink so t hat t he m obile st at ion and UTRAN can com m unicat e wit h t he required QoS. The release of an RRC connect ion can be init iat ed by a request from higher layers or by t he RRC layer it self in case of RRC connect ion failure. I n case of connect ion loss, t he RRC m ay re- est ablish t he RRC connect ion. I n case of RRC connect ion failure, t he RRC releases resources associat ed wit h t he RRC connect ion. Signals t he allocat ion of radio resources t o t he m obile st at ion. Evaluat es, decides, and execut es RRC connect ion m obilit y funct ions such as handover, preparat ion of handover t o GSM or ot her syst em s, cell reselect ion, and cell/ paging area updat e procedures. Ensures and cont rols t hat request ed QoS param et ers are m et . I n t his capacit y, t he RRC allocat es a sufficient num ber of radio resources. Cont rols m easurem ent s done by t he m obile st at ion—t hat is, it decides what t o m easure, when t o m easure, and how t o report . I t also report s t hese m easurem ent s t o t he net work. Select s and reselect s a cell based on idle m ode m easurem ent s and cell select ion crit eria. Broadcast s inform at ion provided by t he net work t o all m obile st at ions. For exam ple, RRC m ay broadcast locat ion service area inform at ion relat ed t o som e specific cells. The RRC configures t he BMC for cell broadcast services and allocat es resources for t his broadcast ing. Perform s ot her funct ions relat ed t o power cont rol, encrypt ion, and int egrit y prot ect ion.

The int eract ion bet ween t he m obile st at ion's RRC and t he UTRAN's RRC in t he cont rol plane can be viewed in Figure 11.31. The RRC also cont rols t he radio resource t hrough t he RLC, MAC, and PHY layers.

Figu r e 1 1 .3 1 . RRC I n t e r a ct ion be t w e e n UTRAN a n d t h e M obile St a t ion

11.6.1 RRC States The m obile st at ion can operat e in eit her idle m ode or connect ed m ode. During st art up, a m obile st at ion perform s t he cell search procedure by scanning for t he broadcast channel ( BCH) . Aft er a cell is discovered and t he m obile st at ion decides t o j oin t he cell ( som et im es called " cam p on a cell" ) , t he m obile st at ion will be in idle m ode and will keep list ening t o t he BCH. I n order t o com m unicat e wit h t he UTRAN, t he m obile st at ion is required t o est ablish an RRC connect ion wit h t he UTRAN and changes t o connect ed m ode. One m obile st at ion can have at m ost one RRC connect ion. As shown in Figure 11.32, in connect ed m ode, t he RRC is in one of t he following four st at es: CELL_DCH, CELL_FACH, URA_PCH, and CELL_PCH. Each RRC st at e reflect s t he physical channels allocat ed t o t he m obile st at ion.

Figu r e 1 1 .3 2 . M obile St a t ion M ode s a n d RRC St a t e s

I n CELL_DCH RRC st at e, t he dedicat ed channel is allocat ed. The RRC 1) reads syst em inform at ion broadcast ed on FACH, 2) perform s m easurem ent s, 3) select s and configures t he radio links and m ult iplexing opt ions applicable for t he t ransport channels, and 4) act s upon RRC m essages received on t he cont rol channels. I n CELL_FACH RRC st at e, t he RACH and FACH are allocat ed. The m obile st at ion 1) m aint ains upt o- dat e syst em inform at ion as broadcast ed by t he serving cell, 2) perform s cell reselect ion processes, 3) perform s m easurem ent s, 4) select s and configures radio links and m ult iplexing opt ions applicable for t he t ransport channels, and 5) act s upon RRC m essages received on t he cont rol channels. When in t his st at e and t he st at ion is out of service, t he RRC perform s a cell select ion process. I n t he URA_PCH or CELL_PCH st at e, t here is no physical channel allocat ed. The m obile st at ion m aint ains up- t o- dat e syst em inform at ion as broadcast ed by t he serving cell, perform s cell reselect ion process, and perform s a periodic search for a higher priorit y net work. I t also m onit ors t he paging channels and perform s m easurem ent s. When t he st at ion is out of service, t he st at ion conduct s a cell select ion process.

Figure 11.33 illust rat es RRC st at es for a m obile st at ion t hat can operat e in dual m ode ( UMTS and GPRS) .

Figu r e 1 1 .3 3 . RRC St a t e s for a D u a l M ode UM TS/ GPRS M obile St a t ion

11.6.2 RRC Procedures The st andard defines several RRC funct ions and procedures. I n t his subsect ion, we int roduce only t he connect ion est ablishm ent and t he radio bearer est ablishm ent procedures.

11.6.2.1 RRC Connection Establishment Procedure The RRC connect ion est ablishm ent procedure ( see Figure 11.34) is init iat ed when t he upper layers in t he m obile st at ion need t o est ablish a signaling connect ion and t he m obile st at ion is in idle m ode. First , t he m obile st at ion's RRC sends an " RRC Connect ion Request " m essage t o UTRAN's RRC t hrough CCCH/ RACH. Then, UTRAN's RRC replies wit h " RRC Connect ion Set up" m essage t hrough CCCH/ FACH. I f t he RRC Connect ion Set up m essage includes a dedicat ed channel assignm ent , t he m obile st at ion will be in connect ed m ode in CELL_DCH st at e. On t he ot her hand, if t he RRC Connect ion Set up m essage includes a com m on channel assignm ent , t he m obile st at ion will be in connect ed m ode in CELL_FACH st at e. Finally, t he m obile st at ion confirm s wit h a " RRC Connect ion Set up Com plet e" m essage t hrough DCCH.

Figu r e 1 1 .3 4 . RRC Con n e ct ion Est a blish m e n t Pr oce du r e

11.6.2.2 Radio Bearer Establishment Procedure Figure 11.35 illust rat es t he radio bearer est ablishm ent procedure. The radio bearer est ablishm ent procedure is init iat ed when t he upper layers in t he m obile st at ion need t o set up t he t ransport and physical channels t o accom m odat e t he t raffic flow from t he upper layers. This procedure st art s aft er t he UTRAN perform s t he adm ission cont rol and select s t he param et ers for t he t ransport ( i.e., t ransport form at set ) and physical ( i.e., channelizat ion coding) channels.

Figu r e 1 1 .3 5 . Ra dio Be a r e r Est a blish m e n t Pr oce du r e

< Day Day Up >

< Day Day Up >

11.7 QoS Support UMTS services adhere t o t he following QoS principles: ●

●

● ● ● ●

Provide a finit e set of QoS definit ions and at t ribut es t hat can be cont rolled. Their com plexit y should be reasonably low and t he inform at ion involved should be kept low as well. Map bet ween applicat ion requirem ent s and UMTS services. The m apping should consider asym m et ric uplink and downlink. Work wit h current QoS schem es and provide different QoS levels. Support session based QoS and allow m ult iple QoS st ream s per address. Manage QoS t o yield efficient resource ut ilizat ion. Modify QoS at t ribut es when t he session is ongoing and act ive.

The end- t o- end QoS delivered t o t he user m ay be specified by Service Level Agreem ent s ( SLAs) bet ween dom ains, part s of t he net work, and operat ors. The st andard QoS end- t o- end archit ect ure is described in Figure 11.36. The bearer service defines t he m echanism s ( i.e., signaling, user plane t ransport , QoS m anagem ent ) t hat provide QoS support . Each service bearer relies on t he service provided by t he bearer service of t he layer below. To enable QoS m anagem ent based on SLAs across dom ains, considerat ion m ust be given t o 1) t he available bandwidt h and QoS classes and how t hey m ap across dom ains, 2) policing and shaping requirem ent s, 3) securit y m echanism s, and 4) report s of service usage, billing inform at ion, com m ercial inform at ion, and financial obligat ions in case of breech of cont ract , et c.

Figu r e 1 1 .3 6 . UM TS QoS Ar ch it e ct u r e

The st andard provides an overview of t he funct ionalit y needed t o est ablish, m odify, and m aint ain a UMTS link wit h a specific QoS. This funct ionalit y is divided int o cont rol and user planes. The st andard provides exam ples of how such m odules int eract t o request and com m it QoS resources in conj unct ion wit h prot ocols such as Resource Reservat ion Prot ocol ( RSVP) . However, t he st andard does not provide t he algorit hm s necessary t o im plem ent such funct ionalit y.

11.7.1 Control Plane QoS Management Functions As shown in Figure 11.37, t he cont rol plane QoS m anagem ent funct ions include several funct ions t hat are responsible for m anaging, t ranslat ing, adm it t ing, and cont rolling user request s and net work resources.

Figu r e 1 1 .3 7 . QoS M a n a ge m e n t Fu n ct ion s in t h e Con t r ol Pla n e . N S = n e t w or k se r ie s, Ph = Ph ysica l, RAB ( Ra dio Acce ss Be a r e r ) .

The following funct ions in t he cont rol plane provide QoS m anagem ent : ●

●

●

Service m anager coordinat es t he est ablishm ent , m odificat ion, and m aint enance of services. I t provides t he user plane wit h QoS m anagem ent funct ions. I t also signals wit h peer service m anagers and uses services provided by ot her inst ances. The service m anager m ay perform an at t ribut e t ranslat ion t o request lower layer services and ot her cont rol funct ions for service provisioning. Translat ion funct ion m aps bet ween t he int ernal service prim it ives for UMTS and ot her ext ernal prot ocols for service cont rol. The t ranslat ion includes t he conversion bet ween UMTS services and QoS at t ribut es of t he ext ernal net work's service cont rol prot ocol. Adm ission/ capabilit y cont rol m aint ains inform at ion about all t he net work's available resources and all resources allocat ed t o UMTS services. I t det erm ines for each UMTS service request or m odificat ion whet her t he required resources can be provided. When

●

t hese resources are available t hey will be reserved for t his service. Subscript ion cont rol checks t he adm inist rat ive right s of t he user t o request t he service wit h specified QoS at t ribut es.

11.7.2 User Plane QoS Management Functions The user plane QoS m anagem ent funct ions are responsible for QoS signaling and m onit oring of user dat a t raffic. Such funct ions will ensure t hat t he t raffic is delivered wit hin cert ain lim it s im posed by specific QoS at t ribut es as negot iat ed wit h t he UMTS net work. ●

●

●

●

Mapping funct ion provides packet s wit h t he specific inform at ion required t o receive t he int ended QoS at t ribut es. Classificat ion funct ion assigns packet s t o t he est ablished services of t he m obile st at ion according t o t he relat ed QoS at t ribut es. Resource m anager dist ribut es t he available resources bet ween all services sharing t he sam e resource according t o t he required QoS. For exam ple, t he resource m anager can em ploy t echniques such as scheduling, bandwidt h m anagem ent , and radio link power cont rol. Traffic condit ioner is a m odule t hat provides t he t raffic conform ance wit h t he negot iat ed QoS at t ribut es. To achieve t his goal, t he t raffic condit ioner deploys t raffic policing and/ or t raffic shaping m echanism s t o all t raffic. The t raffic policing m echanism m onit ors t he t raffic's act ual QoS at t ribut es and com pares t hem t o t he negot iat ed QoS at t ribut es. I f t hese at t ribut es do not m at ch, t he t raffic policing m echanism m arks t he packet s as nonconform ant or drops t hem . The t raffic shaping m echanism shapes t raffic according t o t he negot iat ed QoS at t ribut es. For a m ore det ailed descript ion of t raffic policing and t raffic shaping m echanism s see Chapt er 3.

An exam ple of dat a flow wit hin t hese various user plane QoS funct ions is shown in Figure 11.38.

Figu r e 1 1 .3 8 . UM TS Use r Pla n e QoS M a n a ge m e n t Fu n ct ion s

11.7.3 QoS Classes UMTS defines four different QoS classes: Conversat ional, St ream ing, I nt eract ive, and Background. The m ain difference bet ween t hese QoS classes is t heir sensit ivit y t o t im e delay. The conversat ional class, for exam ple t elephony, is t he m ost t im e delay sensit ive while t he background class is t he least t im e delay sensit ive. The conversat ional and st ream ing classes carry real- t im e t raffic flows. The conversat ional class is used for applicat ions such as t elephony ( e.g., GSM) , voice over I P and video conferencing applicat ions. Real- t im e conversat ion is always perform ed bet ween peers ( or groups) of live ( hum an) end users. Since t he m axim um t ransfer delay is dict at ed by t he hum an percept ion of video and audio conversat ion, t he lim it for accept able t ransfer delay is very st rict . Failure t o provide low enough t ransfer delay will result in unaccept able low qualit y. The st ream ing class includes applicat ions such as real- t im e video and audio st ream s. The highest accept able delay variat ion over t he t ransm ission m edia is given by t he capabilit y of t he t im e alignm ent funct ion of t he applicat ion. The int eract ive class is m ainly used by applicat ions t hat request dat a from rem ot e servers such as web browsing, dat a ret rieval, and server access. I nt eract ive t raffic is charact erized by t he request - response t im e delay pat t ern of t he end user. The background class is used by applicat ions perform ed in t he background, such as downloads of em ails and files. Since t hese applicat ions are delay insensit ive, t he overall expect at ion of t im e delay is m inim al or nonexist ent . As shown in Table 11.3, each t raffic class is described by a set of QoS param et ers.

Ta ble 1 1 .3 . UM TS Be a r e r Se r vice At t r ibu t e s UM TS Be a r e r Se r vice At t r ibu t e

Con ve r sa t ion a l Cla ss

St r e a m in g Cla ss

I n t e r a ct ive Cla ss

Ba ck gr ou n d Cla ss

Maxim um bit rat e ( kbps) — upper lim it of t he bit rat e wit h which t he UMTS delivers packet s

2048

2048

< 2048— Overhead

< 2048— Overhead

Guarant eed bit rat e ( kbps) — guarant eed num ber of bit s delivered by UMTS wit hin a period of t im e

< 2048

< 2048

Delivery order— indicat es whet her t he UMTS link should provide insequence packet delivery or not

Yes/ No

Yes/ No

Yes/ No

Yes/ No

Maxim um packet size ( oct et s)

1500 or 1502

1500 or 1502

1500 or 1502

1500 or 1502

Delivery of erroneous packet s — indicat es whet her packet s det ect ed as erroneous are delivered or discarded

Yes/ No

Yes/ No

Yes/ No

Yes/ No

Residual BER— indicat es t he undet ect ed bit error rat io in t he delivered packet s

5* 10 - 2 , 10 - 2 , 5* 10 3 , 10 - 3 , 10 - 4 , 10 - 5 , 10 - 6

5* 10 - 2 , 10 - 2 , 5* 10 - 3 , 10 - 3 , 10 - 4 , 10 - 5 , 10 -

4* 10 - 3 , 10 - 5 , 6* 10 - 8

4* 10 - 3 , 10 - 5 , 6* 10 - 8

Packet error rat io — indicat es t he fract ion of packet s lost or det ect ed as erroneous

10 - 2 , 7* 10 - 3 , 10 - 3 , 10 - 4 , 10 - 5

10 - 1 , 10 - 2 , 7* 10 - 3 , 10 - 3 , 10 - 4 , 10 - 5

10 - 3 , 10 - 4 , 10 - 6

10 - 3 , 10 - 4 , 10 - 6

Transfer delay ( m s) — indicat es t he m axim um delay for 95 t h percent ile of delay dist ribut ion for all delivered packet s during t he lifet im e of t he com m unicat ion link

100— Maxim um value

250— Maxim um value

6

Under work by 3GPP

Traffic handling priorit y— different iat es and priorit izes bet ween t raffic wit h different QoS requirem ent s wit hin a t raffic class Allocat ion/ Ret ent ion priorit y — different iat es and priorit izes bet ween t raffic classes Source st at ist ic descript or— specifies source charact erist ics SDU form at inform at ion— indicat es t he fract ion of packet s lost or det ect ed as erroneous

Under work by 3GPP

Under work by 3GPP

Speech/ unknown

Speech/ unknown

Under work by 3GPP

Under work by 3GPP

Under work by 3GPP

Under work by 3GPP

11.7.4 QoS Mechanisms in the Radio Bearer Service Figure 11.39 depict s a sim plified diagram of t he radio bearer key QoS m echanism s t hat reside in t he RRC, MAC, and PHY layers. The RRC, which plays t he cent er role for QoS m anagem ent , m anages t he signaling bet ween t he m obile st at ion and t he UTRAN. Furt herm ore, t he RRC can cont rol t he MAC and PHY layers.

Figu r e 1 1 .3 9 . QoS M e ch a n ism s in Ra dio Be a r e r

First , an applicat ion perform s QoS negot iat ion wit h t he UMTS core net work in t he upper layer. The upper layer service at t ribut es are m apped int o t he radio access bearer ( RAB) service at t ribut es. The RRCs of t he m obile st at ion and UTRAN perform t he radio bearer set up and reconfigurat ion procedures. During t hese procedures t he UTRAN's RRC consult s wit h UTRAN's adm ission cont rol m odule in order t o check if t here are enough resources and m ake sure t hat t he new connect ion does not int erfere and does not degrade t he exist ing services. I n case t he applicat ion is adm it t ed, t he RRC allocat es bandwidt h t hrough channelizat ion code assignm ent in t he physical layer and t ransport form at com binat ion in t he MAC layer. Furt herm ore, t he MAC also cont ains a priorit y handling m echanism . Next , we describe t he following aspect s of t he radio bearer QoS m echanism s: bandwidt h allocat ion and priorit y handling.

11.7.4.1 Bandwidth Allocation Aft er a new applicat ion perform s t he QoS negot iat ion, t he RRC est ablishes t he radio bearer services which define t he t ransport and physical channels. The t ransport channel is cont rolled by t he MAC while t he physical channel is cont rolled by t he PHY. As discussed in Sect ion 11.3.1, t he param et ers t hat describe t he t ransport channel are included in t he Transport Form at : t ransport block size, t ransport block set size, and t ransport t im e int erval. Therefore, t hese param et ers will define t he bandwidt h allocat ion.

The physical channel param et er is t he channelizat ion code. As described in Sect ion 11.3, t he bandwidt h allocat ion can be changed by changing t he spread fact or ( SF) . The sm aller t he spread fact or code, t he larger t he bandwidt h ( or dat a rat e) . The queue size of t he t ransport layer ( at t he uplink and at t he downlink) is used t o det erm ine t he applicat ion's dynam ic bandwidt h dem and. When t his queue grows beyond a cert ain t hreshold, t he RRC will perform radio bearer reconfigurat ion by adj ust ing t he physical channel configurat ion ( i.e., adj ust channelizat ion code) and t he t ransport channel configurat ion ( i.e., adj ust t ransport form at and channelizat ion code) . Figures 11.40 t o 11.42 show such reconfigurat ion exam ples for different scenarios.

Figu r e 1 1 .4 0 . I n cr e a se of Uplin k Ba n dw idt h Usin g Ph ysica l Ch a n n e l Re con figu r a t ion

Figu r e 1 1 .4 2 . I n cr e a se of Uplin k Ba n dw idt h Usin g Tr a n spor t Ch a n n e l Re con figu r a t ion

Figu r e 1 1 .4 1 . I n cr e a se of D ow n lin k Ba n dw idt h Usin g Ph ysica l Ch a n n e l Re con figu r a t ion

11.7.4.2 Priority Handling Priorit y handing is locat ed in t he MAC- c/ sh as shown in Figure 11.30. The t raffic wit h t he highest priorit y receives service first . The priorit y value is relat ed t o t he Transport Form at select ion. A high priorit y value is associat ed wit h a Transport Form at t hat provides high dat a rat e while a low priorit y value is associat ed wit h a Transport Form at t hat provides low dat a rat e. < Day Day Up >

< Day Day Up >

Chapter 12. cdma2000 Sect ion 12.1. I nt roduct ion Sect ion 12.2. cdm a2000 Archit ect ure Sect ion 12.3. Physical Layer Sect ion 12.4. Media Access Cont rol ( MAC) Sect ion 12.5. Link Access Cont rol ( LAC) Sect ion 12.6. QoS Support < Day Day Up >

< Day Day Up >

12.1 Introduction cdm a2000 is a 3G t echnology whose st andards are specified by t he Third Generat ion Part nership Proj ect 2 ( 3GPP2) . 3GPP2 is a collaborat ive organizat ion t hat includes m em bers from all over t he world, wit h a m aj orit y of m em bers from Nort h Am erica and Asia. 3GPP2 is part of I MT- 2000 ( I nt ernat ional Mobile Telecom m unicat ion–2000) effort t hat is designed t o provide high- speed com m unicat ion wit h high- qualit y m ult im edia services and global roam ing support . Figure 12.1 shows t he evolut ion of t he cdm a2000 st andard. cdm aOne, or I S- 95, was influenced by various st andards such as Advanced Mobile Phone Service ( AMPS) , I S- 136 TDMA, and GSM. I S- 95 has t wo revisions: I S- 95A and I S- 95B. I S- 95A is a 2G t echnology wit h a dat a rat e of 14.4 kbps, while I S- 95B is 2.5G t echnology wit h a dat a rat e of 115 kbps. The 3G cdm a2000 cont inued t o evolve and support I S- 95.

Figu r e 1 2 .1 . cdm a 2 0 0 0 Evolu t ion

cdm a2000, which is also referred t o as I MT2000- Mc ( I MT- CDMA Mult icarrier) , has been divided int o t he following t wo phases: Phase 1 cdm a2000 1x ( som et im es called cdm a2000 1xRTT) and Phase 2 cdm a2000 3x ( som et im es called cdm a2000 3xRTT) . cdm a2000 1x deploys a single radio frequency carrier ( 1.25 MHz bandwidt h) and delivers 307 kbps in a m obile environm ent , whereas cdm a2000 3x deploys m ult icarrier t echnology ( i.e., a m ult iple of 1.25 MHz bandwidt h carriers) and delivers speeds of up t o 2 Mbps. cdm a2000 1x cont inued t o evolve t o cdm a2000 1x Evolut ion ( 1xEV) . cdm a2000 1xEV is backwards com pat ible wit h cdm a2000 1x and cdm aOne. cdm a2000 1xEV ut ilizes Frequency Division Duplex ( FDD) , where t he uplink ( from t he m obile st at ion t o t he base st at ion, also called reverse link) and downlink ( from t he base st at ion t o t he m obile st at ion, also called forward link) operat e on t wo separat e frequency bands. cdm a2000 1xEV has t wo variat ions: cdm a2000 1xEV- DO ( Dat a Only) opt im ized for high- speed

dat a t ransm ission ( up t o 2.4 Mbps) and cdm a2000 1xEV- DV ( Dat a and Voice) , which support s bot h dat a and voice wit h speeds of up t o 3.09 Mbps. cdm a2000 has been int roduced in several releases. I n t his book we describe cdm a2000 Release C. We focus on t he following 3GPP2 docum ent s: ●

●

●

●

●

3GPP2 C.S0001- C I nt roduct ion t o cdm a2000 St andards for Spread Spect rum Syst em s Release C 3GPP2 C.S0002- C Physical Layer St andard for cdm a2000 Spread Spect rum Syst em s Release C 3GPP2 C.S0003- C Medium Access Cont rol ( MAC) St andard for cdm a2000 Spread Spect rum Syst em s Release C 3GPP2 C.S0004- C Signaling Link Access Cont rol ( LAC) St andard for cdm a2000 Spread Spect rum Syst em s Release C 3GPP2 C.S0005- C Upper Layer ( Layer 3) Signaling St andard for cdm a2000 Spread Spect rum Syst em s Release C < Day Day Up >

< Day Day Up >

12.2 cdma2000 Architecture Figure 12.2 illust rat es t he basic net work archit ect ure of cdm aOne ( I S- 95) and cdm a2000. A m obile st at ion connect s t o t he ext ernal net work ( i.e., Public Swit ched Telephone Net work [ PSTN] or I nt ernet ) t hrough t he base st at ion and core net work. The base st at ion consist s of t wo ent it ies: Base Transceiver St at ion ( BTS) and Base St at ion Cont roller ( BSC) . BTS provides t he com m unicat ion services wit hin it s coverage area or cell. BSC m anages t he call handoff and radio resources of each BTS. I n t he I S- 95 syst em , t he BSC connect s t o t he Mobile Swit ching Cent er ( MSC) accom m odat ing bot h voice and dat a t raffic. The voice t raffic is rout ed t hrough t he MSC t o t he ext ernal t elephone net works ( i.e., PSTN) , whereas t he dat a t raffic is rout ed t o t he ext ernal dat a net work ( i.e., I nt ernet ) t hrough t he I nt erworking Funct ion ( I WF) . I WF provides t he access point t o t he I nt ernet . I n a cdm a2000 syst em , t here are separat e links bet ween t he base st at ion and t he core net work. The link bet ween t he BSC and t he MSC accom m odat es voice services while t he link bet ween t he BSC and t he Packet Dat a Service Node ( PDSN) accom m odat es dat a services. PDSN support s, est ablishes, m aint ains, and t erm inat es I P sessions for t he m obile st at ion. The AAA m odule is responsible for aut horizat ion and account ing.

Figu r e 1 2 .2 . cdm a On e a n d cdm a 2 0 0 0 N e t w or k Ar ch it e ct u r e

12.2.1 cdma2000 Air Interface Protocol Architecture Figure 12.3 illust rat es cdm a2000 air int erface prot ocol archit ect ure. The air int erface prot ocol archit ect ure is divided int o layers corresponding t o t he Open Syst em s I nt erconnect ion ( OSI ) prot ocol layers: 1) Layer 1 or physical layer; 2) Layer 2, which includes t he Link Access Cont rol ( LAC) sublayer, t he Media Access Cont rol ( MAC) sublayer, and t he Forward Packet Dat a Channel ( F- PDCH) Cont rol Funct ion sublayer; and 3) Layers 3 t o 7, which include t he upper layer signaling and t he voice and dat a services originat ed from t he users' applicat ions. The FPDCH Cont rol Funct ion m odule newly int roduced in Revision C cont ains several key feat ures such as Adapt ive Modulat ion and Coding schem e ( AMC) and Hybrid Aut om at ic Repeat request ( HARQ) . There are also signaling int erfaces bet ween t he upper layer and t he physical layer.

Figu r e 1 2 .3 . cdm a 2 0 0 0 Air I n t e r fa ce Pr ot ocol Ar ch it e ct u r e

A m ore det ailed descript ion of cdm a2000 air int erface prot ocol archit ect ure at t he m obile st at ion side is shown in Figure 12.4.

Figu r e 1 2 .4 . D e t a ile d cdm a 2 0 0 0 Air I n t e r fa ce Pr ot ocol Ar ch it e ct u r e ( M obile St a t ion Side )

cdm a2000 is based on physical and logical channels t hat carry bot h dat a and cont rol packet s. The logical channel defines what t ype of inform at ion ( i.e., dat a m essage, cont rol m essage) is delivered. Packet s from each logical channel are m ult iplexed and delivered in t im ely fashion by t he MAC t o t he appropriat e physical channels. The MAC delivers t hese packet s based on t he

QoS requirem ent s of each logical channel. The physical channel defines t he radio t ransport channel based on t he radio configurat ion, which is a com binat ion of encoding, int erleaving, fram e size, and bit rat e. The radio configurat ion of t he physical channel is defined in t he physical layer. As shown in Figure 12.4 t he connect ion bet ween t he MAC and t he physical layers indicat es t he physical channel t o which t he connect ion belongs. There are t wo direct ions on bot h t he logical and physical channels: forward direct ion ( from t he base st at ion t o t he m obile st at ion) and reverse direct ion ( from t he m obile st at ion t o t he base st at ion) . These direct ions are included as t he first let t er ( f = forward, r = reverse) of t he logical and physical channel nam e. A logical channel nam e consist s of t hree lower case let t ers and t he suffix " ch" ( channel) . The first let t er ( i.e., f = forward and r = reverse) indicat es t he direct ion of t he logical channel. " f" and " r" can be used t oget her if t he logical channel has bot h direct ions. The second let t er ( i.e., d = dedicat ed or c = com m on) indicat es t he t ype of logical channel ( dedicat e or com m on [ share] ) . The t hird let t er ( i.e., t = t raffic or s = signaling) indicat es t he cont ent s. The first and second let t ers are separat ed by a hyphen. Exam ples of logical channels include f- dt ch ( forward dedicat ed t raffic channel) and f/ r- dsch ( forward/ reverse dedicat ed signaling channel) . The physical channel nam e is writ t en in capit al let t ers. The first let t er ( i.e., F = forward or R = reverse) indicat es t he direct ion of t he physical channel such as F- PDCH ( Forward Packet Dat a Channel) . < Day Day Up >

< Day Day Up >

12.3 Physical Layer cdm a2000 uses an FDD Code Division Mult iple Access ( CDMA) net work—t hat is, t he t ransm ission from t he base st at ion t o t he m obile st at ion, referred t o as forward t raffic ( i.e., downlink) , is done on a different frequency t han t hat of t he t raffic from t he m obile st at ion t o t he base st at ion, referred t o as reverse t raffic ( i.e., uplink) as shown in Figure 12.5. For each forward and reverse link, cdm a2000 em ploys a com binat ion of Tim e Division Mult iplex ( TDM) and Code Division Mult iplex ( CDM) . For cdm a2000 1x, t he frequency channel widt h of bot h t he forward and reverse links is 1.25 MHz. These frequency channels are separat ed by 45 MHz.

Figu r e 1 2 .5 . Ch a n n e l Acce ss

As shown in Table 12.1, cdm a2000 is assigned various forward and reverse frequencies based on t he count ry of applicat ion.

Ta ble 1 2 .1 . cdm a 2 0 0 0 Fr e qu e n cy Ba n ds Tr a n sm it Fr e qu e n cy Ba n d ( M H z)

Ba n d Cla ss

Syst e m

M obile St a t ion ( Re ve r se Lin k )

Ba se St a t ion ( For w a r d Lin k )

0

Nort h Am erica Cellular

824- 849

869- 894

1

Nort h Am erica PCS

1850- 1910

1930- 1990

2

Tot al Access Com m unicat ion Syst em

872- 915

917- 960

3

Japan Tot al Access Com m unicat ion Syst em

887- 925

832- 870

4

Korean PCS

1750- 1780

1840- 1870

5

Nordic Mobile Telephone

411- 484

421- 494

6

I MT- 2000

1920- 1980

2110- 2170

7

Nort h Am erica 700

776- 794

746- 764

8

1800 MHz

1710- 1785

1805- 1880

9

900 MHz

880- 915

925- 960

The physical layer delivers t he packet s received from t he MAC layer t o t he physical channels. The physical layer defines a com binat ion of fram es, m odulat ions, and codes used for each physical channel. For channel access, t he physical channel deploys CDMA, in which t he inform at ion bit s are spread by using a num ber of Pseudo- Noise ( PN) chip signals ( a m ult iple of 1.2288 Mcps) . The cdm a2000 Release C defines t wo spreading rat es: Spreading Rat e 1 and Spreading Rat e 3. Spreading Rat e 1 ( referred t o as 1x) uses 1.2288 Mcps PN chip signal spread over a 1.25 MHz channel. Spreading Rat e 3 ( referred t o as 3x) has t wo spreading approaches. The first approach uses 1.2288 Mcps PN chip signal spreading over each of t hree 1.25 MHz channels ( som et im es called Mult icarrier Channel) . This approach is used in forward CDMA channels. The ot her approach uses 3.6864 Mcps PN chip signal spreading over a 3.75 MHz channel. The lat t er approach is used in reverse CDMA channels. Figure 12.6 illust rat es t he Spreading Rat e.

Figu r e 1 2 .6 . Spr e a din g Ra t e

The CDMA involves t wo t ypes of codes: long code and Walsh code. The long code ( sim ilar t o t he scram bling code in Universal Mobile Telecom m unicat ions Syst em [ UMTS] ) uses PN sequence t o scram ble t he forward and reverse CDMA channels. On t he reverse channel, different long codes are used t o ident ify t ransm issions from different m obile st at ions. On t he forward channel, t he phase of t he long code is used t o ident ify t he base st at ion t ransm issions. Unlike coding in UMTS, where t he different scram bling codes are used t o different iat e bet ween t ransm issions from different base st at ions, in cdm a2000 each base st at ion uses t he sam e long code but different phases. There are a t ot al of 512 long code phases. The Walsh code ( sim ilar t o t he channelizat ion code in UMTS) is used t o ident ify t he physical channels wit hin a m obile st at ion or wit hin a base st at ion. cdm a2000 has a t ot al of 128 Walsh codes. The basic concept of Walsh code, which is referred t o as channelizat ion code, is described in Chapt er 11, Sect ion 11.3. cdm a2000 also defines t he radio configurat ions ( RCs) for t he forward and reverse CDMA channels. The radio configurat ion consist s of a com binat ion of physical layer param et ers such as spreading m odulat ion ( i.e., BPSK, QPSK) , spreading rat e ( i.e., 1x, 3x) , forward error correct ion ( i.e., 1/ 2 t o 1/ 4 convolut ion code, 1/ 2 t o 1/ 5 t urbo code) , and dat a rat e. There are six radio configurat ions ( RC1 t o RC6) for t he reverse CDMA channel and t en radio configurat ions ( RC1 t o RC10) for t he forward CDMA channel.

12.3.1 Physical Channels The st andard defines t he st ruct ure of t he physical channel in fram es. The fram e st ruct ure varies on each t ype of physical channel wit h possible fram e sizes of 1.25 m s, 2.5 m s, 5 m s, 10 m s, 20 m s, 40 m s, and 80 m s. The physical channels include t he following:

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

Forward/ Reverse Fundam ent al Channel ( F/ R- FCH) Forward/ Reverse Dedicat ed Cont rol Channel ( F/ R- DCCH) Forward/ Reverse Supplem ent al Code Channel ( F/ R- SCCH) Forward/ Reverse Supplem ent al Channel ( F/ R- SCH) Paging Channel ( F- PCH) Quick Paging Channel ( F- QPCH) Access Channel ( R- ACH) Forward/ Reverse Com m on Cont rol Channel ( F/ R- CCCH) Forward/ Reverse Pilot Channel ( F/ R- PI CH) Transm it Diversit y Pilot Channel ( F- TDPI CH) Auxiliary Pilot Channel ( F- API CH) Auxiliary Transm it Diversit y Pilot Channel ( F- ATDPI CH) Sync Channel ( F- SYNCH) Com m on Power Cont rol Channel ( F- CPCCH) Com m on Assignm ent Channel ( F- CACH) Enhanced Access Channel ( R- EACH) Broadcast Cont rol Channel ( F- BCCH) Forward Packet Dat a Channel ( F- PDCH) Forward Packet Dat a Cont rol Channel ( F- PDCCH) Reverse Acknowledgm ent Channel ( R- ACKCH) Reverse Channel Qualit y I ndicat or Channel ( R- CQI CH)

I n t he following subsect ion we int roduce som e of t hese physical channels.

12.3.1.1 Pilot Channels There are pilot channels in bot h forward and reverse links. I n t he forward link, pilot channels include Forward Pilot Channel ( F- PI CH) , Transm it Diversit y Pilot Channel ( F- TDPI CH) , Auxiliary Pilot Channel ( F- API CH) and Auxiliary Transm it Diversit y Pilot Channel ( F- ATDPI CH) . Forward Pilot Channel is t ransm it t ed by t he base st at ion at all t im es. F- PI CH helps t he m obile st at ion in t he init ial cell search process. When a new m obile st at ion j oins a cell, it det ect s t he F- PI CH t ransm it t ed by t he base st at ion. The m obile st at ion also m easures t he forward signal st rengt h from t he pilot channel. F- TDPI CH has t he sam e funct ion as F- PI CH but it is used in t ransm it diversit y. F- API CH and F- ATDPI CH are used in beam form ing ant enna syst em s. The Reverse Pilot Channel ( R- PI CH) is only a pilot channel in t he reverse link which helps t he base st at ion det ect t he m obile st at ion. R- PI CH also includes t he Reverse Power Cont rol Subchannnel which cont rols t he power of t he forward link.

12.3.1.2 Physical Channels for Voice and Data Traffic Figure 12.7, which is a part of Figure 12.4, illust rat es t he physical channels for voice and dat a t raffic. cdm a2000 1x defines Forward/ Reverse Fundam ent al Channel ( F/ R- FCH) , Forward/ Reverse Supplem ent al Channel ( F/ R- SCH) , and Forward/ Reverse Dedicat ed Cont rol Channel ( F/ R- DCCH) t o support voice and dat a t raffic. I n addit ion, cdm a2000 1xEV int roduces t he F- PDCH Cont rol Funct ion, which cont ains new physical channels t hat support t he high- speed forward dat a t ransm ission. These new physical channels include Forward Packet Dat a Channel ( FPDCH) , Forward Packet Dat a Cont rol Channel ( F- PDCCH) , Reverse Acknowledgm ent Channel ( RACKCH) , and Reverse Channel Qualit y I ndicat or Channel ( R- CQI CH) .

Figu r e 1 2 .7 . Ph ysica l Ch a n n e ls for Voice a n d D a t a Tr a ffic ( m obile st a t ion side )

The Forward/ Reverse Fundam ent al Channel ( F/ R- FCH) can support voice, dat a, and signaling. The st andard defines a flexible dat a rat e t hat ranges from 750 bps t o 14.4 kbps by varying t he spreading rat e and t he associat ed set of fram es for each radio configurat ion. The Forward/ Reverse Dedicat ed Cont rol Channel ( F/ R- DCCH) can be used for signaling or burst dat a t ransm issions. The Forward/ Reverse Supplem ent al Channel ( F/ R- SCH) is defined t o support high rat e dat a services. F/ R- SCH is scheduled dynam ically on a fram e- by- fram e basis. F/ R- SCH can deliver dat a rat es up t o 32 t im es t he rat es of an F/ R- FCH. At t he base st at ion, t he num ber of F- SCHs is lim it ed by t he available t ransm ission power and Walsh codes. At t he m obile st at ion, t he num ber of R- SCHs is lim it ed t o t wo. Forward Packet Dat a Channel ( F- PDCH) provides high- speed forward dat a t ransm ission. A base st at ion can support up t o t wo F- PDCHs, and each F- PDCH t ransm it s dat a packet s t o one m obile st at ion at a t im e. A m obile st at ion can support one F- PDCH at a t im e. Forward Packet Dat a Cont rol Channel ( F- PDCCH) provides t he t ransm it t ing cont rol inform at ion ( i.e., users' MAC I D, packet size, subpacket I D) for t he associat ed F- PDCH. F- PDCCH can also be used t o broadcast t he Walsh m ask t o t he m obile st at ions. A base st at ion can support up t o t wo F- PDCCHs ( corresponding t o t wo F- PDCHs) . When a m obile st at ion det ect s it s own MAC I D on a F- PDCCH, it will ret rieve dat a packet s from t he associat ed F- PDCH. The high forward dat a rat e can be achieved t hrough Adapt ive Modulat ion and Coding ( AMC) schem e and flexible Tim e Division Mult iplex/ Code Division Mult iplex ( TDM/ CDM) . The AMC schem e includes a com binat ion of a num ber of inform at ion bit s ( i.e., 386, 770, 1538, 2306, 3074 or 3842) , t he t ot al fram e durat ion ( i.e., 1.25 m s., 2.5 m s, or 5 m s) , m odulat ion schem es ( i.e., QPSK, 8- PSK or 16- QAM) , and spreading codes. Table 12.2 shows F- DPCH possible dat a rat es.

Ta ble 1 2 .2 . F- D PCH D a t a Ra t e s ( k bps)

F- PD CH Pa ck e t Size ( bit s) ( in for m a t ion bit s + 1 6 qu a lit y in dica t or bit s + 6 t u r bo e n code r t a il bit s)

Tot al Fram e Durat ion ( m s)

408 bit s

792 bit s

1 5 6 0 bit s

2 3 2 8 bit s

3 0 9 6 bit s

3 8 6 4 bit s

5 ms

82 kbps

158 kbps

312 kbps

466 kbps

619 kbps

773 kbps

2.5 m s

163 kbps

317 kbps

624 kbps

931 kbps

1238 kbps

1546 kbps

1.25 m s 326 kbps

634 kbps

1248 kbps

1862 kbps

2477 kbps

3091 kbps

The AMC schem e has flexible choices t o achieve t he different dat a rat es. TDM/ CDM provides packet scheduling of m ult iple F- PDCHs ( up t o t wo F- PDCHs t ransm it t ing at t he sam e t im e) by varying t he Walsh Code and scheduling. An exam ple of a TDM/ CDM assignm ent is shown in Figure 12.8 for four m obile st at ions ( MS1 t o MS4) .

Figu r e 1 2 .8 . Ex a m ple of a TD M / CD M Assign m e n t

The Reverse Acknowledgm ent Channel ( R- ACKCH) enables Aut om at ic Repeat Request ( ARQ) m echanism . Aft er successfully receiving packet s from t he base st at ion, t he m obile st at ion sends an acknowledgm ent t o t he base st at ion t hrough t he R- ACKCH. The fact t hat t he ARQ m echanism resides at t he physical layer speeds up t he ret ransm ission process. Furt herm ore, inst ead of using t he t ypical ARQ m echanism , cdm a2000 1xEV em ploys a Hybrid ARQ t hat com bines ARQ wit h Forward Error Correct ion ( FEC) . To im prove t he error correct ion perform ance and reduce t he num ber of ret ransm ission at t em pt s, Hybrid ARQ perform s FEC decoding on bot h erroneous packet s from t he previous t ransm ission and packet s from t he current t ransm ission. Reverse Channel Qualit y I ndicat or Channel ( R- CQI CH) is used by t he m obile st at ion t o indicat e t o t he base st at ion t he channel qualit y m easurem ent s. The base st at ion t hen uses t his inform at ion t o det erm ine t he appropriat e spreading, m odulat ion, fram e size, and scheduling of F- PDCH. I t allows t he base st at ion t o efficient ly m anage t he radio resources. I n sum m ary, t he voice and dat a packet s are t ransm it t ed from t he base st at ion t hrough F- FCH, F- DCCH, F- SCH, and F- PDCH ( in case high- speed forward t ransm ission is required) . The voice and dat a packet s are t ransm it t ed from t he m obile st at ion t o t he base st at ion t hrough F- FCH, FDCCH, and F- SCH.

< Day Day Up >

< Day Day Up >

12.4 Media Access Control (MAC) Figure 12.9 illust rat es t he MAC layer and it s int eract ions wit h t he LAC. cdm a2000 support s a generalized m odel of m ult im edia services t hat allows sim ult aneous support for bot h voice and dat a t raffic. cdm a2000 also includes Qualit y of Service ( QoS) cont rol m echanism s t hat reside at t he MAC layer. These m echanism s balance different QoS requirem ent s of m ult iple sim ult aneous m ult im edia applicat ions. The key funct ions of t he MAC layer are 1) t o receive t he packet s ( i.e., voice, dat a) from t he upper layers and 2) t o schedule t hese packet s t o physical channels in a t im ely m anner based on t he connect ion QoS requirem ent s. This process is done by t he MAC t hrough a Mult iplexing and QoS Delivery m odule ( see Figure 12.9) . Release C st andard also defines t he signaling bet ween t he LAC and t he MAC using service prim it ives. This signaling provides QoS inform at ion t hat can be used by t he MAC. The st andard also defines t he following t wo m odules: Signaling Radio Burst Prot ocol ( SRBP) and Radio Link Prot ocol ( RLP) . SRBP is a connect ionless prot ocol for signaling m essages. RLP, which is a connect ion- orient ed prot ocol, provides reasonably reliable t ransm ission over t he radio link using a negat ive- acknowledgm ent based dat a delivery prot ocol.

Figu r e 1 2 .9 . M AC La ye r

We int roduce four key service prim it ives used in t he signaling procedure bet ween t he Signaling LAC and t he MAC:

●

MAC- SDUReady.Request : This prim it ive is sent from t he signaling LAC t o t he MAC when t here are packet s wait ing in t he signaling LAC. The prim it ive includes t he following param et ers: - channel_t ype: t he t ype of channel required—" 5m s FCH/ DCCH fram e," " 20m s FCH/ DCCH fram e," or " F- PDCH fram e" - size: t he size of packet s ( bit s) - scheduling hint : indicat es t he relat ive priorit y service requirem ent

●

MAC- Availiabilit y.I ndicat ion: This prim it ive is sent from t he MAC t o t he signaling LAC when t he MAC m ult iplex m odule is ready t o receive t he packet s from t he upper layer and t o t ransm it t hem t o t he physical layer. The prim it ive includes t he following param et ers: - channel_t ype: t he t ype of channel allowed—" 5m s FCH/ DCCH fram e," " 20m s FCH/ DCCH fram e," or " F- PDCH fram e" - m ax_size: t he m axim um num ber of bit s from t he signaling LAC t hat can be fit t ed in t he physical channel based on QoS requirem ent s - syst em _t im e: indicat es t he t im e when t he physical layer will t ransm it t he packet s

●

●

MAC- Dat a.Request : This prim it ive is sent from t he signaling LAC t o t he MAC. I t includes t he t ype of channel required and a packet . The MAC m ult iplex m odule will fit t his packet t o t he closest size dat a block det erm ined by t he rat e set of t he physical channel. MAC- Dat a.I ndicat ion: This prim it ive is sent from t he MAC t o t he signaling LAC t o indicat e t he packet t ransm ission in t he physical layer.

Figure 12.10 illust rat es t he signaling and packet t ransm ission process ( st eps 1 t o 5) at t he MAC layer.

Figu r e 1 2 .1 0 . Sign a lin g a n d Pa ck e t Tr a n sm ission Pr oce ss

12.4.1 Logical Channels The following logical channels define what t ype of inform at ion ( i.e., user dat a m essage, cont rol m essage) is delivered: ●

●

●

Forward/ Reverse Dedicat ed Traffic Channel ( f/ r- dt ch) : A point - t o- point logical channel t hat carries voice and dat a packet s and t ransm it s t hrough a dedicat ed physical channel. Forward/ Reverse Dedicat ed Signaling Channel ( f/ r- dsch) : A point - t o- point logical channel t hat carries signaling packet s from t he upper layer t o a dedicat ed physical channel. Forward/ Reverse Com m on Signaling Channel ( f/ r- csch) : A point - t o- m ult ipoint logical channel t hat carries signaling packet s from t he upper layer t o a com m on physical channel.

12.4.2 Multiplex and QoS Sublayers The m ult iplex sublayer is responsible for t ransm it t ing and receiving packet s from t he physical

layer. The m ult iplex sublayer receives inform at ion from t he logical channels t hat com es from various sources ( i.e., upper layer signaling, dat a, and voice services) . The t ransm it t ing funct ion, under QoS cont rol, solicit s inform at ion from various packet s exchanged wit h t he MAC. This inform at ion is used by t he m ult iplex sublayer t o det erm ine t he relat ive priorit y bet ween t he t raffic supplied by signaling and ot her services. The exact m anner for using t his inform at ion t o deliver over- t he- air QoS is not specified by t he cdm a2000 st andard. < Day Day Up >

< Day Day Up >

12.5 Link Access Control (LAC) The LAC, which includes several sublayers as depict ed in Figure 12.11, perform s t he following funct ions: ● ●

● ●

●

Delivery of packet s wit h opt ional ARQ t echniques t o provide reliabilit y Segm ent at ion of packet s t o sizes suit able for t he MAC and reassem bly of such packet s com ing from t he MAC Access cont rol t hrough aut hent icat ion Address cont rol t o ensure delivery of packet s based upon addresses t hat ident ify part icular m obile st at ions I nt ernal signaling, by exchanging not ificat ions and dat a wit h t he MAC and t he supervisory and configurat ion ent it ies, result ing from t he processing of LAC Sublayer level inform at ion

Figu r e 1 2 .1 1 . LAC Su bla ye r

< Day Day Up >

< Day Day Up >

12.6 QoS Support cdm a2000 Release C st andard support s voice and high- speed dat a t ransm ission. Each applicat ion has different QoS requirem ent s. As we described in previous sect ions, cdm a2000 defines several QoS m echanism s t hat reside at different prot ocol layers ( i.e., physical layer, MAC layer, and upper layers) . I n t his sect ion we sum m arize t he QoS m echanism s defined by t he st andard.

12.6.1 Bandwidth Allocation The st andard defines Adapt ive Modulat ion and Coding ( AMC) and flexible TDM/ CDM. Bot h t echniques provide flexible t ools t hat enable dynam ic bandwidt h allocat ion t o m at ch wit h t he current t raffic and channel condit ions. The algorit hm t hat det erm ines t he bandwidt h allocat ion is not defined by t he st andard. The product developers need t o develop t hese algorit hm s in order t o opt im ize t heir net work resources.

12.6.2 Packet Scheduling Packet scheduling algorit hm s indicat e when packet s from a user or applicat ion are allowed t o t ransm it . The st andard defines t he m ult iplex and QoS delivery sublayers in t he MAC t o support packet scheduling algorit hm s. Mult iplex and QoS delivery sublayers can provide service different iat ion am ong logical channels ( i.e., voice, dat a, signaling) . The st andard also defines t he service prim it ives bet ween t he signaling LAC and MAC t hat enable t he t ransfer of QoS requirem ent inform at ion from t he upper layers. The packet scheduling algorit hm s are not defined by t he st andard. < Day Day Up >

< Day Day Up >

Chapter 13. Satellite Communication Sect ion 13.1. I nt roduct ion Sect ion 13.2. Archit ect ure Sect ion 13.3. Forward Link Sect ion 13.4. Ret urn Link Sect ion 13.5. Qualit y of Service Support < Day Day Up >

< Day Day Up >

13.1 Introduction Com m unicat ion sat ellit es allow t he est ablishm ent of com m unicat ion worldwide in bot h populat ed and rem ot e areas, where ot her ways of com m unicat ion m ay be im possible ( e.g., oceans and m ount ains) . Tradit ionally, sat ellit es have been used for broadcast ing t elevision program s. I n recent years, t he sat ellit e indust ry has evolved and has st art ed t o provide t wo- way com m unicat ion. This evolut ion is possible due t o t echnology, decreasing cost of user's equipm ent , and it s availabilit y in all world regions. These sat ellit e syst em s are eit her propriet ary or st andardized. Hughes Net work Syst em s wit h it s Spaceway program and Gilat Sat ellit e Net works are exam ples of propriet ary syst em s. I n Europe, ETSI ( European Telecom m unicat ions St andards I nst it ut e) has been developing st andards for sat ellit e com m unicat ion. The sat ellit e st andards are published under t he DVB ( Digit al Video Broadcast ing) proj ect . ETSI is cooperat ing wit h t he DVB consort ium of hundreds of broadcast ers, m anufact urers, net work operat ors, soft ware developers, regulat ory bodies, and ot hers in m any count ries com m it t ed t o designing global st andards for t he delivery of digit al t elevision and dat a services. Com panies such as SESASTRA, Eut elsat , and Hispasat have been using t he DVB st andard. Sat ellit es can orbit t he eart h in several orbit s referred t o as GEO ( Geosynchronous, or Geost at ionary Eart h Orbit ) , MEO ( Medium , or Middle Eart h Orbit ) , and LEO ( Low Eart h Orbit ) . These sat ellit es work in t he Ku band ( 10–17 GHz frequency range) or t he Ka band ( 18–31GHz frequency range) . GEO sat ellit es orbit t he eart h at 22,300 m iles above t he eart h's surface. They are t ied t o t he eart h's rot at ion and t heir posit ions are fixed in relat ion t o eart h's surface. This is an advant age since t he eart h st at ion needs t o point it s t ransm it t er and receiver t o only one locat ion in space t o be able t o t ransm it and receive from t he GEO sat ellit e. This m akes GEO sat ellit es popular for t ransm issions of high- speed dat a, t elevision, and ot her wideband applicat ions. A disadvant age is t he long dist ance t hat t he signal needs t o t ravel. This result s in a long t im e delay of few hundred m illisecond, which causes det eriorat ion in t he QoS support provided for int eract ive applicat ions t hat require short delays. MEO sat ellit es orbit t he eart h bet ween 1,000 and 22,300 m iles above t he eart h's surface. They are not st at ionary in relat ion t o eart h's surface. MEO sat ellit es are used in geographical posit ioning syst em s. LEO sat ellit es orbit t he eart h bet ween 400 and 1,000 m iles above t he eart h's surface. LEOs are m ost ly used for dat a com m unicat ion such as paging, em ail, and videoconferencing. LEO sat ellit es do not have a fixed locat ion in space in relat ion t o t he eart h's surface. They m ove at very high speeds and t o sust ain end- t o- end com m unicat ion, t hey need t o com m unicat e wit h each ot her. Because t hey are locat ed relat ively close t o eart h, t he eart h st at ion t ransm it t er uses significant ly less power. Also t he t ravel t im e for t he signal is significant ly short er, a fact t hat significant ly decreases t he m ult im edia applicat ion's delay, pot ent ially im proving QoS support .

13.1.1 DVB Return Channel System (DVB-RCS) The DVB fam ily of st andards cont ains GEO sat ellit e com m unicat ion wit h various opt ions of

uplink ( or ret urn) and downlink ( or forward) channels. Up t o a few years ago, sat ellit es were only able t o t ransfer dat a unidirect ionally from t he cont ent sources t o t he end user equipm ent , m ost ly TV set s. However, in recent years bidirect ional com m unicat ion has been int roduced wit h various ret urn channels such as t elephones, I SDN, Cable TV, DECT, and LMDS. Here is a list of DVB st andards: ●

●

●

●

●

●

●

● ●

●

ETS 300 802, Digit al Video Broadcast ing ( DVB) ; Net work- independent prot ocols for DVB int eract ive services ETSI ES 200 800, Digit al Video Broadcast ing ( DVB) ; DVB int eract ion channel for Cable TV dist ribut ion syst em s ( CATV) TR 101 201, Technical Report Digit al Video Broadcast ing ( DVB) ; I nt eract ion channel for Sat ellit e Mast er Ant enna TV ( SMATV) dist ribut ion syst em s; Guidelines for versions based on sat ellit e and coaxial syst em s EN 301 193, European St andard ( Telecom m unicat ions series) Digit al Video Broadcast ing ( DVB) ; I nt eract ion channel t hrough t he Digit al Enhanced Cordless Telecom m unicat ions ( DECT) EN 301 195, European St andard ( Telecom m unicat ions series) Digit al Video Broadcast ing ( DVB) ; I nt eract ion channel t hrough t he Global Syst em for Mobile com m unicat ions ( GSM) EN 301 199, European St andard ( Telecom m unicat ions series) Digit al Video Broadcast ing ( DVB) ; I nt eract ion channel for Local Mult i- point Dist ribut ion Syst em s ( LMDS) ETS 300 801, Digit al Video Broadcast ing ( DVB) ; I nt eract ion channel t hrough Public Swit ched Telecom m unicat ions Net work ( PSTN) / I nt egrat ed Services Digit al Net works ( I SDN) ETSI EN 301 958 V1.1.1 ( 2002- 03) European St andard ( Telecom m unicat ions series) Digit al Video Broadcast ing ( DVB) ; I nt eract ion channel for Digit al Terrest rial Television ( RCT) incorporat ing Mult iple Access OFDM TR 101 194, Digit al Video Broadcast ing ( DVB) ; Guidelines for im plem ent at ion and usage of t he specificat ion of net work- independent prot ocols for DVB int eract ive services

I n t his book we focus on t he sat ellit e DVB st andards t hat provide t wo- way com m unicat ion t hrough t he sat ellit e, referred t o as DVB- RCS. I n t hese syst em s t he end user has direct access t o t he sat ellit e using t he sat ellit e's uplink and downlink channels. We focus on t he following st andards and docum ent s: ●

●

●

●

●

ETSI EN 301 790, European St andard ( Telecom m unicat ions series) , Digit al Video Broadcast ing ( DVB) ; I nt eract ion channel for sat ellit e dist ribut ion syst em s ETSI TR 101 790, European St andard ( Telecom m unicat ions series) , Technical Report Digit al Video Broadcast ing ( DVB) ; I nt eract ion channel for Sat ellit e Dist ribut ion Syst em s; Guidelines for t he use of EN 301 790 TR 101 202, European St andard ( Telecom m unicat ions series) , Digit al Video Broadcast ing ( DVB) ; I m plem ent at ion guidelines for Dat a Broadcast ing EN 301 192, European St andard ( Telecom m unicat ions series) , Digit al Video Broadcast ing ( DVB) ; DVB specificat ion for dat a broadcast ing ETS 300 802, European St andard ( Telecom m unicat ions series) , Digit al Video Broadcast ing ( DVB) ; Net work- independent prot ocols for DVB int eract ive services

The downlink relat ed part of t he specificat ions includes t he DVB- Sat ellit e ( DVB- S) st andard which is used in t he t ypical sat ellit e broadcast syst em s. For uplink com m unicat ion, t he DVBRCS st andard uses t he RCST ( Ret urn Channel Sat ellit e Term inal) , a sat ellit e t erm inal t hat support s int eract ive services such as SMATV ( Sat ellit e Mast er Ant enna Television) as well as dat a t ransm ission. DVB- RCS MAC is based upon FDM/ TDM, where t he uplink and downlink t ransm issions t ake place at different frequency ranges. DVB- RCS does not define a specific operat ing frequency

channel t o be used, which allows flexible frequency deploym ent . Bot h t he uplink and downlink channels are t im e slot t ed. The MAC specificat ions allow vendor specific im plem ent at ions of various QoS support ing algorit hm s. Depending on t he dish size, DVB- RCS can achieve m axim um dat a speeds as det ailed in Table 13.1. The act ual dat a speed depends on t he DVB operat or equipm ent and agreem ent s wit h t he end user.

Ta ble 1 3 .1 . D VB- RCS M a x im u m D a t a Spe e ds D ish Size

0 .6 M e t e r

1 .0 M e t e r

1 .2 M e t e r

Ret urn Link Speed

150 kbps

380 kbps

2 Mbps

Forward Link Speed

38 Mbps

38 Mbps

38 Mbps

< Day Day Up >

< Day Day Up >

13.2 Architecture The DVB- RCS m odel is shown in Figure 13.1. DVB support s t wo channels: t he Broadcast Channel and t he I nt eract ion Channel.

Figu r e 1 3 .1 . D VB- RCS Syst e m M ode l

The Broadcast Channel, which is a unidirect ional downlink broadcast channel, is ident ical t o t he channel defined in t he sat ellit e digit al video broadcast ( DVB- S) st andard. The I nt eract ion Channel provides bidirect ional int eract ion com m unicat ion bet ween t he service provider and t he end- user. The I nt eract ion Channel consist s of a Forward I nt eract ion channel ( from t he service provider t o end- user) and Ret urn I nt eract ion channel ( from t he end- user t o service provide) . Typically, t he Forward I nt eract ion channel is included in t he Broadcast Channel. The RCST provides int erface for bot h Broadcast and I nt eract ion Channels. I t is support ed by t he Net work I nt erface Unit and t he Set Top Unit . A sat ellit e int eract ive net work wit h several RCST channels is described in Figure 13.2. The NCC ( Net work Cont rol Cent er) m onit ors and cont rols t he operat ion of t he sat ellit e int eract ive net work. The NCC m anages t he net work resources and aut horizes and allocat es t ransm ission resources t o RCSTs. The syst em is fed by a few Feeder St at ions. The Feeder St at ion on t he forward link is a st andard sat ellit e digit al video broadcast ( DVB- S) on which we m ult iplex user dat a, cont rol, and t im ing signals. The Traffic Gat eway receives RCST ret urn signals and provides account ing funct ions, int eract ive services, and connect ions t o ext ernal public or privat e service providers. These service providers can be I nt ernet dat a services, pay- per- view TV, financial services, and corporat e net works.

Figu r e 1 3 .2 . Ove r a ll Sa t e llit e I n t e r a ct ive N e t w or k Ar ch it e ct u r e

The forward link carries signaling inform at ion from t he NCC and user t raffic t o RCSTs. The signaling t raffic from t he NCC t o RCSTs, referred t o as t he Forward Link Signaling, includes t he necessary inform at ion required t o operat e t he ret urn link. Bot h t he user t raffic and t he forward link signaling can be carried over different forward link signals. Several RCST configurat ions are possible depending on t he num ber of forward link receivers present on t he RCST. I n t he next sect ions we describe t he forward and reverse link t echnologies t hat are used in t he DVB- RCS st andard. < Day Day Up >

< Day Day Up >

13.3 Forward Link The forward link or t he broadcast channel is based on t he DVB- S st andard which em ploys MPEG2 Transport St ream ( TS) defined by I SO/ I EC ( I nt ernat ional St andards Organizat ion/ I nt ernat ional Elect rot echnical Com m ission) 13818- 1 st andard. Mult iple dat a st ream s ( i.e., video, audio, dat a) are m ult iplexed on MPEG- 2 TS and delivered t hrough t he broadcast channel as shown in Figure 13.3. These dat a st ream s are cat egorized int o several applicat ion areas as follows: DVB dat a piping, DVB dat a st ream ing, DVB m ult iprot ocol encapsulat ion, DVB dat a carousel, and DVB obj ect carousel. Each applicat ion area has different QoS requirem ent s. Som e dat a st ream s ( i.e., DVB dat a piping, DVB dat a st ream ) are m apped direct ly ont o MPEG- TS. Ot her dat a st ream s ( i.e., DVB m ult iprot ocol encapsulat ion, DVB dat a carousel, and DVB obj ect carousel) are first m apped t hrough Dat a St orage Media Com m and and Cont rol ( DSM- CC) and t hen MPEG- TS. The Forward Link em ploys Tim e Division Mult iplex ( TDM) on a single digit al carrier. Before describing each applicat ion area, we would like t o provide som e background of MPEG- 2 Transport St ream .

Figu r e 1 3 .3 . D a t a Br oa dca st Spe cifica t ion Ove r vie w ( sa t e llit e side )

13.3.1 MPEG-2 Transport Stream MPEG- 2 Transport St ream is a t ransport form at for video and audio t ransm ission. Figure 13.4 illust rat es a sim plified overview of t he MPEG- 2 process.

Figu r e 1 3 .4 . M PEG- 2 Pr oce ss Ove r vie w

The video encoder com presses t he video signal and const ruct s video fram es ( i.e., I - fram e, Pfram e, B- fram e) . Each video fram e includes t he elem ent st ream ( ES) header and a series of encoded video fram es. The ES header provides inform at ion about t he cont ent and locat ion of t he video blocks required for decoding t he video st ream . The elem ent st ream s are packet ized and const ruct ed int o a packet ized elem ent st ream ( PES) . Figure 13.5 shows t he PES packet st ruct ure. The PES header cont ains t he inform at ion necessary t o decode t he video st ream . I m port ant inform at ion in PES header t hat we would like t o point out includes: 1) t he Decode Tim e St am p ( DTS) , which indicat es when t o decode t he video fram e and 2) t he Present at ion Tim e St am p ( PTS) , which indicat es when t o display t he video fram e. PES packet s from t he different st ream s are m ult iplexed int o a Transport St ream ( TS) . Figure 13.6 shows t he Transport St ream packet st ruct ure. The packet I D ( PI D) in t he TS header ident ifies t he TS payload cont ent s.

Figu r e 1 3 .5 . PES Pa ck e t St r u ct u r e

Figu r e 1 3 .6 . Tr a n spor t St r e a m Pa ck e t St r u ct u r e

The list of PI D and it s associat ed values are included in t he Program Specific I nform at ion ( PSI ) t able. There are several t ypes of PSI t ables such as t he Program Associat ion Table ( PAT) , t he

Program Map Table ( PMT) , and t he Condit ional Access Table ( CAT) . Figure 13.7 illust rat es t he PSI t able st ruct ure. Wit hin t he MPEG- 2 Transport St ream , t here are m ult iple program s ( i.e., TV program s) . Each program consist s of m ult iple video st ream s, audio st ream s, and dat a st ream s. The PSI t able provides t he PI D inform at ion required t o find, ident ify, and reconst ruct t he video st ream in subsequent Transport St ream packet s. First , t he receiver scans for t he PAT ident ified by a reserved PI D ( PI D = 0) . PAT cont ains t he m apping bet ween t he PI D value and t he associat ed program point er of t he PMT. For exam ple, as shown in Figure 13.7, PI D = 22 point s t o t he PMT of program 1. The PMT cont ains t he m apping bet ween t he PI D value and t he associat ed st ream of t he program . For exam ple, as shown in Figure 13.7, PI D = 62 ident ifies t he video st ream of program 1.

Figu r e 1 3 .7 . Pr ogr a m Spe cific I n for m a t ion ( PSI ) Ta ble St r u ct u r e

There are ot her t ypes of t ables such as t he Superfram e Com posit ion Table ( SCT) , Fram e Com posit ion Table ( FCT) , and Tim e Slot Com posit ion Table ( TCT) used t o define t he channel st ruct ure. We will describe t hem lat er in t his chapt er.

13.3.2 Data Piping The dat a piping m echanism uses an asynchronous ( i.e., wit hout t im ing, such as I P packet s) t ransport m echanism . As shown in Figure 13.8, t he dat a packet s are insert ed direct ly int o t he payload of t he MPEG2 t ransport packet s wit hout any fragm ent at ion and reassem bly. At t he TS header, t he payload_unit _st art _indicat or field m ay be im plem ent ed t o include t he indicat ion of t he st art of t he dat a piping packet while t he t ransport _priorit y field m ay be im plem ent ed t o include t he end of t he packet .

Figu r e 1 3 .8 . D a t a Pipin g Pa ck e t s

13.3.3 Data Streaming Dat a st ream ing m echanism is for st ream ing- orient ed applicat ions which m ay be asynchronous, synchronous, or synchronized dat a st ream s. Asynchronous dat a st ream ing t ransm it s dat a wit hout t im ing requirem ent s while synchronous dat a st ream ing t ransm it s dat a using a fixed rat e t ransm ission clock ( i.e., T1, circuit em ulat ion) . Synchronized dat a st ream ing t ransm it s dat a wit h t im ing relat ed t o t he different kinds of dat a st ream s ( i.e., st ream ing video along wit h audio where bot h require t im e synchronizat ion) . As shown in Figure 13.9, t he dat a st ream ing packet s are insert ed int o t he PES packet payload.

Figu r e 1 3 .9 . Pa ck e t ize d Ele m e n t St r e a m ( PES) Pa ck e t s for Asyn ch r on ou s/ Syn ch r on ou s/ Syn ch r on ize d D a t a St r e a m in g

The st ream I D in t he PES packet ident ifies t he t ype of dat a st ream ing cont ained in t he payload. The Present at ion Tim e St am p ( PTS) described in Sect ion 13.3.1 is an opt ional field in t he PES header t hat includes t im e st am p inform at ion for synchronous and synchronized dat a st ream ing.

Packet scheduling algorit hm s can use t he PTS inform at ion t o det erm ine when t o t ransm it synchronous/ synchronized packet s. For synchronous dat a packet s, t here is a field in t he PES payload called out put dat a rat e t hat includes t he synchronous dat a st ream required dat a rat e.

13.3.4 Multiprotocol Encapsulation Mult iprot ocol encapsulat ion m et hod is used for t ransport ing packet s t hat originat e from different net work prot ocols on t op of t he MPEG2 Transport St ream s in DVB net works. I t has been opt im ized for delivery of I P packet s but can be used for ot her prot ocols as well. A 48- bit MAC address is used t o ident ify t he receiver ( i.e., RCST) . UDP/ I P or TCP/ I P t raffic will be m ult iplexed int o an MPEG- 2 t ransport st ream t hrough DSM- CC sect ions ( defined in I SO/ I EC 13818- 6) as det ailed in Figure 13.10. Using encrypt ion, encapsulat ion can also provide secure t ransm ission.

Figu r e 1 3 .1 0 . M u lit ipr ot ocol En ca psu la t ion for I P Tr a ffic

The DSM- CC ( Digit al St orage Media Com m and and Cont rol) facilit at es t he t ransm ission of a st ruct ured group of obj ect s from a broadcast server t o client s using direct ory obj ect s, file obj ect s, and st ream obj ect s. I t provides t ools for cont rolling t he MPEG- 2 st ream s. DSM- CC m ay be used for cont rolling t he video recept ion, providing feat ures norm ally found on Video Casset t e Recorders ( VCRs) such as fast - forward, rewind, and pause. DSM- CC is designed for light weight and fast operat ion considering lim it ed m em ory devices. The long delay im posed by t he dist ances of t housands of m iles t hat t he signal needs t o t ravel adversely affect s t he TCP feat ures such as congest ion cont rol and packet acknowledgm ent s. TCP's congest ion cont rol m echanism assum es t hat all dat a loss is due t o congest ion. Therefore, it wrongly t ranslat es t he long delays t o dat a loss and t hus slows down t he t ransm ission of packet s and prevent s new connect ions. Moreover, t he required TCP acknowledgm ent s can congest t he relat ive sm all bandwidt h available on t he uplink channel. Consequent ly, sat ellit e vendors em ploy solut ions t hat overcom e such TCP problem s. Such solut ions m ay em ploy an int erm ediat e server t hat act s as t he end- user and proj ect s t im ing param et ers t o t he t raffic source such t hat it will not slow down it s t ransm issions. The int erm ediat e server will encapsulat e t he packet s in " new" packet s and will t ransm it t hem t o t he end- user.

13.3.5 Data Carousel The Dat a Carousel is a m echanism t hat allows a server t o present a set of packet s t o a

program , which is run by a receiver, by cyclically repeat ing t he cont ent s of t he carousel one or m ore t im es. An exam ple is t he Telet ext syst em in which a set of pages is cyclically broadcast ed. < Day Day Up >

< Day Day Up >

13.4 Return Link The ret urn link carries t ransm issions from t he RCSTs t o t he sat ellit e. As shown in Figure 13.11, t he ret urn channel is st ruct ured int o superfram es, fram es, and t im e slot s.

Figu r e 1 3 .1 1 . Su pe r fr a m e s, Fr a m e s, a n d Tim e slot s

A superfram e is defined by a sect ion of frequency range and t im e durat ion of t he ret urn channel. Each superfram e is labeled wit h a superfram e_count er. A superfram e consist s of fram es ident ified by fram e_num bers. A fram e consist s of a t im e slot ident ified by t im eslot _num ber. For channel allocat ion purposes, each t im e slot is uniquely ident ified by Superfram e_id, Superfram e_count er, Fram e_num ber, and Tim eslot _num ber. The superfram e, fram e, and t im e slot st ruct ure of t he sat ellit e syst em as well as t heir ident ificat ions are indicat ed in t he Superfram e Com posit ion Table ( SCT) , t he Fram e Com posit ion Table ( FCT) , and t he Tim e Slot Com posit ion Table ( TCT) , which are broadcast ed by t he sat ellit e.

All RCST t ransm issions are governed and cont rolled by t he NCC. Before t he RCST can send dat a, it needs t o j oin t he net work ( logon) and inform t he NCC about it s configurat ion. The logon process occurs on a frequency channel t hat is defined by t he NCC and t ransm it t ed on t he Forward Link. The Ret urn Link support s four t ypes of t raffic: t raffic ( TRF) , acquisit ion ( ACQ) , synchronizat ion ( SYNC) , and com m on signaling channel ( CSC) . TRF is used for carrying dat a from t he RCST t o t he Gat eway St at ion. This t raffic can include ATM based cells or MPEG- 2 packet s. Synchronizat ion and Acquisit ion t raffic is used t o posit ion RCST t ransm issions during and aft er t he logon process. A SYNC is used by an RCST t o synchronize and send cont rol inform at ion t o t he syst em . An ACQ can be used t o achieve synchronizat ion prior t o operat ional use of t he net work by t he RCST. The CSC packet s are only used by t he RCST t o ident ify it self t o t he syst em during t he logon process, as well as ot her set up inform at ion, such as det erm ining t he frequencies.

13.4.1 Return Link Channel Access Because of t he m ult ipoint - t o- point t ransm issions on t he ret urn link, t he RCSTs em ploy a m ult iple access m echanism denot ed Mult ifrequency Tim e Division Mult iple Access ( MF- TDMA) . MF- TDMA em ploys a com binat ion of FDMA and TDMA. I n every superfram e, t he NCC broadcast s a Term inal Burst Tim e Plan ( TBTP) which indicat es t he frequency, t he bandwidt h, t he st art t im e, and t he durat ion for each RCST t hat is allowed t o t ransm it dat a. There are t wo MF- TDMA schem es: fixed- slot MF- TDMA and dynam ic- slot MF- TDMA. I n fixed- slot MF- TDMA, t he bandwidt h and durat ion of t he t im e slot s assigned t o an RCST are fixed while in dynam ic- slot MF- TDMA, t he bandwidt h and durat ion of t he t im e slot s are variable. Dynam ic- Slot MF- TDMA is opt ional in DVB syst em s. Figure 13.12 illust rat es a channel access assignm ent diagram for a specific RCST.

Figu r e 1 3 .1 2 . Ch a n n e l Acce ss Assign m e n t D ia gr a m for a Spe cific RCST

The channel assignm ent is based on t he applicat ion's bandwidt h dem and and QoS requirem ent s. The channel assignm ent process can be achieved t hrough a request - grant process where t he RCST request s t he bandwidt h and t he NCC grant s t he bandwidt h. As shown in Figure 13.13, t he st andard defines capacit y request s as well as signaling m et hods. The st andard does not define t he algorit hm t hat det erm ines t he bandwidt h allocat ion of each RCST.

Figu r e 1 3 .1 3 . Ca pa cit y Re qu e st - Gr a n t Pr oce ss

Each applicat ion ( ident ified by t he Channel_I D) on a RCST request s bandwidt h individually. The per- connect ion capacit y request is cat egorized int o t he following five cat egories: ●

●

●

●

●

Cont inuous Rat e Assignm ent ( CRA) indicat es t he fixed m inim um bandwidt h requirem ent per superfram e. CRA is suit able for applicat ions t hat require a fixed m inim um guarant ee rat e wit h m inim um delay and delay j it t er such as const ant bit rat e ( CBR) applicat ions ( circuit em ulat ion) . Rat e- Based Dynam ic Capacit y ( RBDC) indicat es t he dynam ic bandwidt h requirem ent per superfram e. The current RBDC request will override all previous RBDC request s. To prevent a hanging capacit y assignm ent , RBDC has a predet erm ined t im e- out period associat ed wit h it . I f t he t im e- out period expires, RBDC is set t o zero. RBDC is suit able for variable bit rat e ( VBR) applicat ions. Pract ically, t he com binat ion of CRA and RBDC can be used. CRA provides t he fixed m inim um bandwidt h while RBDC provides dynam ic bandwidt h on t he t op of t he m inim um bandwidt h. Volum e- Based Dynam ic Capacit y ( VBDC) indicat es t he volum e bandwidt h requirem ent per superfram e. These request s are cum ulat ive ( i.e., each request increm ent s on previous request s) . At t he end of each superfram e, t he cum ulat ive am ount of bandwidt h is deduct ed by t he am ount of bandwidt h already assigned in t he superfram e t o t he specific RCST. VBDC is suit able for delay t olerant applicat ions. Absolut e Volum e- Based Dynam ic Capacit y ( AVBDC) is sim ilar t o VBDC wit h t he difference t hat AVBDC is not cum ulat ive. I t indicat es t he absolut e value of t he volum e bandwidt h. Therefore, t he current AVBDC will override all previous AVBDCs. AVBDC is used in cases where VBDC m ight be lost during t he request process. Free Capacit y Assignm ent ( FCA) uses t he available bandwidt h assignm ent . FCA is not m apped t o any t raffic cat egory. The available bandwidt h assignm ent is considered as bonus bandwidt h used t o reduce t raffic delays.

The st andard defines t he following signaling m et hods t hat can be used for capacit y request s: ●

Sat ellit e Access Cont rol ( SAC) field: A SYNC packet used for synchronizat ion cont ains an opt ional SAC field. This field includes t he connect ion capacit y request . The SYNC packet

●

●

●

can be used in cont ent ion m ode where m ult iple RCSTs t ransm it SYNC packet s at t he sam e t im e. Dat a Unit Labeling Met hod ( DULM) : This m et hod allows RCSTs t o t ransm it cont rol and m anagem ent inform at ion t o t he NCC piggybacked t o t he dat a packet s already sent during t he assigned periods for TRF t raffic. Mini- slot Met hod: This m et hod is based on a periodic assignm ent t o already logged- on RCSTs for packet s t hat use less t han full t raffic t im e slot s. I t carries cont rol and m anagem ent inform at ion and is used also for m aint aining RCST synchronizat ion. This m echanism is support ed by t he SAC used in SYNC packet s. Cont ent ion- based Mini- slot Met hod: This m et hod can be applied such t hat a group of RCSTs can access a Mini- Slot on a cont ent ion basis. This m echanism is support ed by t he SAC request used in SYNC packet s. The access is based on Slot t ed Aloha, in which t he RCST t ransm it s at t he beginning of t he t im e slot and if it is not successful it backs off and ret ransm it s in anot her t im e slot . < Day Day Up >

< Day Day Up >

13.5 Quality of Service Support Figure 13.14 shows DVB- RCS QoS archit ect ure. The st andard defines som e basic QoS m echanism s such as classificat ion, channel access, and capacit y request signaling.

Figu r e 1 3 .1 4 . D VB- RCS QoS Ar ch it e ct u r e

13.5.1 Classification Before packet s from a specific applicat ion can be t ransm it t ed by t he RCST, a connect ion needs t o be est ablished bet ween t he applicat ion and t he NCC. Each applicat ion will be ident ified by t he Channel_I D. DVB- RCS uses per- flow classificat ion which enables per- flow QoS services. The signaling processes involved in t he connect ion set up, m odificat ion, and t erm inat ion are expect ed t o be clarified in t he next revision of t he st andard.

13.5.2 Capacity Request Signaling and Channel Access The st andard defines t he signaling m et hod for capacit y request s as well as t he t ypes of capacit y request s. As described in Sect ion 13.4.1 each t ype of capacit y request which is per connect ion basis, support s different t ypes of applicat ions wit h different QoS requirem ent s.

As described in Sect ion 13.4.1, t he ret urn link em ploys MF- TDMA. Each RCST is allowed t o t ransm it packet s in a specific frequency and for a specific am ount of t im e as defined in t he Term inal Burst Tim e Plan ( TBTP) as described in Sect ion 13.4.1. TBTP is sent from t he NCC t o t he RCST in every superfram e. The TBTP m echanism enables dynam ic bandwidt h allocat ion which can provide QoS support for each applicat ion. The scheduling algorit hm t hat det erm ines t he TBTP as a funct ion of t he applicat ion's QoS requirem ent s as well as t he available syst em resources is not included in t he st andard. < Day Day Up >

< Day Day Up >

Appendix Acronyms and Abbreviations Part 1: Mult im edia Applicat ion and Qualit y of Service ( QoS) Part 2: Wireless Local Area Net works Part 3: Wireless Met ropolit an Area Net works Part 4: Wireless Personal Area Net works Part 5: 2.5G and 3G Net works < Day Day Up >

< Day Day Up >

Part 1: Multimedia Application and Quality of Service (QoS) CBR

Const ant Bit Rat e

CSMA/ CA

Carrier Sense Mult iple Access wit h Collision Avoidance

CSMA/ CD

Carrier Sense Mult iple Access wit h Collision Det ect ion

DiffServ

Different iat ed Service

DSCP

Different ial Service Code Point

EMS

Enhanced Messaging Service

GPS

Global Posit ioning Syst em

I nt Serv

I nt egrat ed Service

MMS

Mult im edia Messaging Service

MOS

Mean Opinion Score

QoS

Qualit y of Service

RSVP

Resource Reservat ion Prot ocol

SBM

Subnet Bandwidt h Manager

SMS

Short Message Service

TDMA

Tim e Division Mult iple Access

TOS

Type Of Service

VBR

Variable Bit Rat e

VOD

Video On Dem and

VoI P

Voice over I P

WAP

Wireless Applicat ion Prot ocol

WFQ

Weight Fair Queue

< Day Day Up >

< Day Day Up >

Part 2: Wireless Local Area Networks ACF

Associat ion Cont rol Funct ion

ACH

Access Feedback Channel

AI FS

Arbit rat ion I nt erfram e Space

AP

Access Point

ARQ

Aut om at ic Repeat Request

ASCH

Associat ion Cont rol Channel

BC

Broadcast

BCCH

Broadcast Cont rol Channel

BCH

Broadcast Channel

BSS

Basic Service Set

CC

Cont rolled Cont ent ion

CCA

Clear Channel Assessm ent

CCI

Cont rolled Cont ent ion I nt erval

CFP

Cont ent ion- Free Period

CM

Cent ralized Mode

CP

Cont ent ion Period, Connect ion Point

CTS

Clear To Send

CW

Cont ent ion Window

DCC

DLC Connect ion Cont rol

DCCH

Dedicat ed Cont rol Channel

DCF

Dist ribut ed Coordinat ion Funct ion

DFS

Dynam ic Frequency Select ion

DI FS

DCF I nt erfram e Space

DiL

Direct Link

DL

DownLink

DLC

Dat a Link Cont rol

DM

Direct Mode

DS

Dist ribut ion Syst em

DSSS

Direct Sequence Spread Spect rum

EC

Error Cont rol

EDCF

Enhanced DCF

EI FS

Ext ended I nt erfram e Space

ESS

Ext ended Service Set

FCCH

Fram e Cont rol Channel

FCH

Fram e Channel

FHSS

Frequency Hopping Spread Spect rum

HC

Hybrid Coordinat or

HCF

Hybrid Coordinat ion Funct ion

I BSS

I ndependent Basic Service Set

I FS

I nt erfram e Space

LCCH

Link Cont rol Channel

LCH Long

Transport Channel

MT

Mobile Term inal

NAV

Net work Allocat ion Vect or

OFDM

Ort hogonal Frequency Division Mult iplexing

PCF

Point Coordinat ion Funct ion

PF

Persist ence Fact or

PI FS

PCF I nt erfram e Space

PLCP

Physical Layer Convergence Prot ocol

QSTA

QoS St at ion

RA

Random Access

RBCH

RLC Broadcast Channel

RCH

Random Access Channel

RFCH

Random Access Feedback Channel

RG

Resource Grant s

RLC

Radio Link Cont rol

RR

Resource Request

RRC

Radio Resource Cont rol

RTS

Request To Send

SCH

Short Transport Channel

SI FS

Short I nt erfram e Space

TC

Traffic Cat egory

TS

Traffic St ream

TSPEC

Traffic Specificat ions

TXOP

Transm ission Opport unit y

UBCH

User Broadcast Channel

UDCH

User Dat a Channel

UL

Uplink

UMCH

User Mult icast Channel

UWB

Ult ra Wide Band

< Day Day Up >

< Day Day Up >

Part 3: Wireless Metropolitan Area Networks BE

Best Effort

BS

Base St at ion

CI D

Connect ion I dent ifier

CPS

Com m on Part Sublayer

CS

Convergence Sublayer

DCD

Downlink Channel Descript or

DI UC

Downlink I nt erval Usage Code

GPC

Grant Per Connect ion

GPSS

Grant Per Subscriber St at ion

IE

I nform at ion Elem ent

I UC

I nt erval Usage Code

nrt PS

Non- Real- Tim e Polling Service

PMP

Point t o Mult ipoint

rt PS

Real- Tim e Polling Service

SI

Slip I ndicat or

SS

Subscriber St at ion

UCD

Uplink Channel Descript or

UGS

Unsolicit ed Grant Service

UI UC

Uplink I nt erval Usage Code

< Day Day Up >

< Day Day Up >

Part 4: Wireless Personal Area Networks ACL

Asynchronous Connect ion- Less

CAC

Channel Access Code

CAP

Cont ent ion Access Period

CFP

Cont ent ion- Free Period

CTA

Channel Tim e Allocat ion

CTR

Channel Tim e Request

DAC

Device Access Code

FHS

Frequency Hop Synchronizat ion

GTS

Guarant eed Tim e Slot

HCI

Host Cont rol I nt erface

HI D

Hum an I nt erface Device

I AC

I nquiry Access Code

L2CAP

Logical Link Cont rol and Adapt at ion Prot ocol

LC

Link Cont rol

LM

Link Manager

PNC

PicoNet Coordinat or

PNI D

PicoNet I D

MTS

Managem ent Tim e Slot

SCO

Synchronous Connect ion- Orient ed

SDP

Service Discovery Prot ocol

UA

User Asynchronous

UI

User I sochronous

US

User Synchronous

< Day Day Up >

< Day Day Up >

Part 5: 2.5G and 3G Networks GPRS, UMTS, cdma2000 3GPP

3G Part nership Proj ect

3GPP2

3G Part nership Proj ect 2

AAA

Aut horizat ion And Account ing

AMC

Adapt ive Modulat ion and Coding

AMR

Adapt ive Mult irat e

ARI B

Associat ion of Radio I ndust ries and Businesses

ARQ

Aut om at ic Repeat Request

ASC

Access Service Classes

BCCH

Broadcast Cont rol Channel

BCH

Broadcast Channel

BMC

Broadcast / Mult icast Cont rol

BSC

Base St at ion Cont roller

BSS

Base St at ion Subsyst em

BTS

Base Transceiver St at ion

CCCH

Com m on Cont rol Channel

CCH

Cont rol Channel

CCI TT

Consult at ive Com m it t ee for I nt ernat ional Telephony and Telegraphy

CDMA

Code Division Mult iple Access

CPAGCH

Com pact Packet Access Grant Channel

CPBCCH

Com pact Packet Broadcast Cont rol Channel

CPCH

Com m on Packet Channel

CPNCH

Com pact Packet Not ificat ion Channel

CPPCH

Com pact Packet Paging Channel

CPRACH

Com pact Packet Random Access Channel

CSCF

Call St at e Cont rol Funct ion

CTCH

Com m on Traffic Channel

CWTS

China Wireless Telecom m unicat ion St andard Group

DBPSCH

Dedicat ed Basic Physical Subchannel

DCCH

Dedicat ed Cont rol Channel

DCH

Dedicat ed Channel

DL

Dat a Link

DSCH

Downlink Shared Channel

DTCH

Dedicat ed Traffic Channel

DTM

Dual Transfer Mode

EDGE

Enhanced Dat a Rat e for Global Evolut ion

EGPRS

Enhanced GPRS

ETSI

European Telecom m unicat ions St andards I nst it ut e

FACH

Forward Access Channel

F- API CH

Forward Dedicat ed Auxiliary Pilot Channel

F- ATDPI CH

Forward Auxiliary Transm it Diversit y Pilot Channel

F- BCCH

Forward Broadcast Cont rol Channel

F- CACH

Forward Com m on Assignm ent Channel

F- CPCCH

Forward Com m on Power Cont rol Channel

F- PCH

Forward Paging Channel

F- PDCCH

Forward Packet Dat a Cont rol Channel

F- PDCH

Forward Packet Dat e Channel

F- QPCH

Forward Quick Paging Channel

FR

Full Rat e

F/ R- CCCH

Forward/ Reverse Com m on Cont rol Channel

F/ R- DCCH

Forward/ Reverse Dedicat ed Cont rol Channel

F/ R- FCH

Forward/ Reverse Fundam ent al Channel

F/ R- PI CH

Forward/ Reverse Pilot Channel

F/ R- SCCH

Forward/ Reverse Supplem ent al Code Channel

F/ R- SCH

Forward/ Reverse Supplem ent al Channel

F- SYNCH

Forward Synchronous Channel

F- TDPI CH

Forward Transm it Diversit y Pilot Channel

GERAN

GSM/ EDGE Radio Access Net work

GGSN

Gat eway GPRS Support Node

GPRS

General Packet Radio Service

G- RNTI

GERAN Radio Net work Tem porary I dent it y

GSM

Global Syst em for Mobile Com m unicat ions

HARQ

Hybrid Aut om at ic Repeat Request

HLR

Hom e Locat ion Regist er

HR

Half Rat e

HS

High- Speed

HSCSD

High- Speed Circuit Swit ched Dat a

HS- DSCH

High- Speed Downlink Shared Channel

HSDPA

High- Speed Downlink Packet Access

HSPDA

High- Speed Download Packet Access

HSS

Hom e Subscriber Server

I MS

I nt ernet Mult im edia Subsyst em

I MT- 2000

I nt ernat ional Mobile Telecom m unicat ion–2000

I TU

I nt ernat ional Telecom m unicat ions Union

LAC

Link Access Cont rol

MAC

Medium Access Cont rol

MGCF

Media Gat eway Cont rol Funct ion

MGW

Media Gat eway

MRF

Mult im edia Resource Funct ion

MS

Mobile St at ion

MSC

Mobile Swit ching Cent er

OVSF

Ort hogonal Variable Spreading Fact or

PACCH

Packet Associat ed Cont rol Channel

PAGCH

Packet Access Grant Channel

PBCCH

Packet Broadcast Cont rol Channel

PCCCH

Packet Com m on Cont rol Channel

PCH

Paging Channel

PCM

Pulse Code Modulat ion

PDCCH

Packet Dedicat ed Cont rol Channel

PDCH

Packet Dat a Channel

PDCHCF

Packet Dat a Channel Cont rol Funct ion

PDCP

Packet Dat a Convergence Prot ocol

PDP

Packet Dat a Prot ocol

PDSN

Packet Dat a Serving Node

PDTCH

Packet Dat a Traffic Channel

PNCH

Packet Not ificat ion Channel

PPCH

Packet Paging Channel

PRACH

Packet Random Access Channel

PSTN

Public Swit ched Telephone Net work

PTCCH/ D

Packet Tim ing Advance Cont rol Channel/ Downlink

PTCCH/ U

Packet Tim ing Advance Cont rol Channel/ Uplink

PTM

Point - To- Mult ipoint

PTM- M

Point - To- Mult ipoint –Mult icast

PTP

Point - To- Point

QoS

Qualit y of Service

RACH

Random Access Channel

R- ACH

Reverse Access Channel

R- ACKCH

Reverse Acknowledgm ent Channel

RAN

Radio Access Net work

RCN

Radio Net work Cont roller

R- CQI CH

Reverse Channel Qualit y I ndicat or Channel

R- EACH

Reverse Enhanced Access Channel

RLC

Radio Link Cont rol

RLP

Radio Link Prot ocol

RNC

Radio Net work Cont roller

RR

Radio Resource

RRC

Radio Resource Cont rol

R- SGW

Roam ing Signaling Gat eway

RX

Receive

SBPSCH

Shared Basic Physical Subchannel

SF

Spread Fact or

SGSN

Serving GPRS Support Node

SHCCH

Shared Channel Cont rol Channel

SRBP

Signaling Radio Burst Prot ocol

SS7

Signaling Syst em No. 7

TBF

Tem porary Block Flow

TCH

Traffic Channel

TDMA

Tim e Division Mult iple Access

TE

Term inal Equipm ent

TFC

Transport Form at Com binat ion

TFCS

Transport Form at Com binat ion Set

TFI

Transport Form at I ndicat or, Tem porary Flow I dent it y

TI A/ EI A

Telecom m unicat ions I ndust ry Alliance/ Elect ronics I ndust ries

TLLI

Tem porary Logical Link I dent ifier

TTA

Telecom m unicat ion Technology Associat ion

TTI

Transm ission Tim e I nt erval

T- SGW

Transport Signaling Gat eway

TX

Transm it

UE

User Equipm ent

UMTS

Universal Mobile Telecom m unicat ions Syst em

USCH

Uplink Shared Channel

USF

Uplink St at e Flag

UTRAN

UMTS Terrest rial Radio Access Net work

VLR

Visit or Locat ion Regist er

WCDMA

Wide CDMA or Wide Code Division Mult iple Access

Satellite Communication ACQ

Acquisit ion

AVBDC

Absolut e Volum e Based Dynam ic Capacit y

CRA

Cont inuous Rat e Assignm ent

CSC

Com m on Signaling Channel

DSM- CC

Digit al St orage Media Com m and and Cont rol

DTS

Decode Tim e St am p

DULM

Dat a Unit Labeling Met hod

DVB

Digit al Video Broadcast ing Proj ect

DVB- S

Digit al Video Broadcast by Sat ellit e

FCA

Free Capacit y Assignm ent

GEO

Geosynchronous or Geost at ionary Eart h Orbit

I SO

I nt ernat ional St andards Organizat ion

LEO

Low Eart h Orbit

MEO

Medium or Middle Eart h Orbit

NCC

Net work Cont rol Cent er

PAT

Program Associat ion Table

PES

Packet ized Elem ent St ream

PI D

Packet I D

PMT

Program Map Table

PSI

Program - Specific I nform at ion

PTS

resent at ion Tim e St am p

QPSK

Quadrat ure Phase Shift Keying

RBDC

Rat e- Based Dynam ic Capacit y

RCST

Ret urn Channel Sat ellit e Term inal

SAC

Sat ellit e Access Cont rol

SMATV

Sat ellit e Mast er Ant enna Television

SYNC

Synchronizat ion

TBTP

Term inal Burst Tim e Plan

TRF

Traffic

TS

Traffic St ream

VBDC

Volum e- Based Dynam ic Capacit y

< Day Day Up >

< Day Day Up >

Bibliography I nt roduct ion Part 1: Mult im edia Applicat ions and Qualit y of Service ( QoS) Part 2: Wireless Local Area Net works Part 3: Wireless Met ropolit an Area Net works Part 4: Wireless Personal Area Net works Part 5: 2.5G and 3G Net works < Day Day Up >

< Day Day Up >

Introduction R. D. Nee and R. Prasad, OFDM for Wireless Mult im edia Com m unicat ions, Art ech House, Norwood, MA, 2000. M. K. Sim on, J. K. Om ura, R. A. Scholt z, and B. K. Levit t , Spread Spect rum Com m unicat ions Handbook, McGraw- Hill Professional, New York, 2001. R. St evens, The Prot ocols ( TCP/ I P I llust rat ed, Volum e 1) , Addison- Wesley, Bost on, 1994. A. Tanenbaum , Com put er Net works, Prent ice Hall, Upper Saddle River, NJ, 2002. < Day Day Up >

< Day Day Up >

Part 1: Multimedia Applications and Quality of Service (QoS) RFC ( Request for Com m ent s, I nt ernet RFC/ STD/ FYI / BCP Archives) 1349, P. Alm quist , " Type of Service in t he I nt ernet Prot ocol Suit e," July 1992. RFC 1633, R. Braden, D. Clark, and S. Shenker, " I nt egrat ed Services in t he I nt ernet Archit ect ure: an Overview," June 1994. RFC 2205, R. Braden, L. Zhang, S. Berson, S. Herzog, and S. Jam in, " Resource ReSerVat ion Prot ocol ( RSVP) ," Sept em ber 1997. RFC 2211, J. Wroclawski, " Specificat ion of t he Cont rolled- Load Net work Elem ent Service," Sept em ber 1997. RFC 2212, S. Shenker, C. Part ridge, and R. Guerin, " Specificat ion of Guarant eed Qualit y of Service," Sept em ber 1997. RFC 2474, K. Nichols, S. Blake, F. Baker, and D. Black, " Definit ion of t he Different iat ed Services Field ( DS Field) in t he I Pv4 and I Pv6 Headers," Decem ber 1998. RFC 2475, S. Blake, D. Black, M. Carlson, E. Davies, Z. Wang, and W. Weiss, " An Archit ect ure for Different iat ed Services," Decem ber 1998. RFC 2814, R. Yavat kar, D. Hoffm an, Y. Bernet , F. Baker, and M. Speer, " SBM ( Subnet Bandwidt h Manager) : A Prot ocol for RSVP- Based Adm ission Cont rol over I EEE 802- St yle Net works," May 2000. Y. Bernet , Net working Qualit y of Service and Windows Operat ing Syst em s, Que Publishing, I ndianapolis, I N, 2001. A. Croll and E. Packm an, Managing Bandwidt h: Deploying QoS in Ent erprise Net works, Prent ice Hall, Upper Saddle River, NJ, 2000. A. Dem ers, S. Keshav, and S. Shenker, " Analysis and Sim ulat ion of a Fair Queuing Algorit hm ," SI GCOMM CCR 19, no. 4 ( 1989) . S. Floyd and V. Jacobson, " Link- Sharing and Resource Managem ent Models for Packet Net works," I EEE/ ACM Transact ions on Net working 3, no. 4, 1995. F. Fluckiger, Underst anding Net worked Mult im edia, Prent ice Hall, Upper Saddle River, NJ, 1995. D. Miras, A Survey on Net work QoS Needs of Advance I nt ernet Applicat ions ( Working Docum ent ) , I nt ernet 2 QoS Working Group, Decem ber 2002. Available at ht t p: / / qos.int ernet 2. edu/ wg/ apps/ fellowship/ Docs/ I nt ernet 2AppsQoSNeeds.ht m l. S. V. Raghavan and S. K., Tripat hi, Net worked Mult im edia Syst em s: Concept s, Archit ect ure,

and Design, Prent ice Hall, Upper Saddle River, NJ, 1998. R. St einm et z, and K. Nahrst edt , Mult im edia: Com put ing, Com m unicat ions and Applicat ions, Prent ice Hall, Upper Saddle River, NJ, 1995. < Day Day Up >

< Day Day Up >

Part 2: Wireless Local Area Networks I EEE 802.11 WG, ANSI / I EEE St d 802.11, " I nform at ion t echnology—Telecom m unicat ions and inform at ion exchange bet ween syst em s—Local and m et ropolit an area net works—Specific Requirem ent s—Part 11: Wireless LAN Medium Access Cont rol ( MAC) and Physical Layer ( PHY) Specificat ions," 1999. I EEE 802.11 WG, I EEE St d 802.11a- 1999, " Supplem ent t o I EEE St andard for I nform at ion Technology— Telecom m unicat ions and inform at ion exchange bet ween syst em s—Local and m et ropolit an area net works—Specific requirem ent s—Part 11: Wireless LAN Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions: High- speed Physical Layer in t he 5 GHz Band," 1999. I EEE 802.11 WG, I EEE St d 802.11b- 1999, " Supplem ent t o I EEE St andard for I nform at ion Technology—Telecom m unicat ions and inform at ion exchange bet ween syst em s—Local and m et ropolit an area net works—Specific requirem ent s—Part 11: Wireless LAN Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions: Higher- speed Physical Layer Ext ension in t he 2.4 GHz Band," 1999. I EEE 802.11 WG, I EEE St d 802.11e/ D3.0," Draft Supplem ent t o STANDARD for Telecom m unicat ions and I nform at ion Exchange Bet ween Syst em s—LAN/ MAN Specific Requirem ent s—Part 11: Wireless Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions: Medium Access Cont rol ( MAC) Enhancem ent s for Qualit y of Service ( QoS) ," May 2002. I EEE 802.11 WG, I EEE St d 802.11g/ D2.1, " Draft Supplem ent t o I EEE St andard for I nform at ion Technology—Telecom m unicat ions and inform at ion exchange bet ween syst em s—Local and m et ropolit an area net works—Specific requirem ent s—Part 11: Wireless LAN Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions: Furt her Higher- speed Physical Layer Ext ension in t he 2.4 GHz Band," January 2002. ETSI BRAN, TR 101 301, " Broadband Radio Access Net work ( BRAN) ; High Perform ance Radio Local Area Net work ( HI PERLAN) Type 2; Requirem ent s and archit ect ure for wireless broadband access," January 1999. ETSI BRAN, ETSI TR 101 683, " Broadband Radio Access Net work ( BRAN) ; HI PERLAN Type 2; Syst em Overview," February 2000. ETSI BRAN, ETSI TR 101 761- 1, " Broadband Radio Access Net work ( BRAN) ; HI PERLAN Type 2; Dat a Link Cont rol ( DLC) Layer; Part 1: Basic Dat a Transport Funct ions," Decem ber 2001. ETSI BRAN, ETSI TR 101 761- 2, " Broadband Radio Access Net work ( BRAN) ; HI PERLAN Type 2; Dat a Link Cont rol ( DLC) Layer; Part 2: Radio Link Cont rol ( RLC) sublayer," January 2002. Hom eRF Technical Com m it t ee, " Hom eRF Specificat ion Revision 2.0 Draft 20010507," May 2001. < Day Day Up >

< Day Day Up >

Part 3: Wireless Metropolitan Area Networks I EEE 802.16 WG, I EEE 802.16- 2001, " I EEE St andard for Local and Met ropolit an Area Net works— Part 16: Air I nt erface for Fixed Broadband Wireless Access Syst em s," April 2002. I EEE 802.16 WG, I EEE P802.16a/ D4- 2002, " Draft Am endm ent t o I EEE St andard for Local and Met ropolit an Area Net works—Part 16: Air I nt erface for Fixed Broadband Wireless Access Syst em s—Medium Access Cont rol Modificat ions and Addit ional Physical Layer Specificat ions for 2- 11 GHz," May 2002. Eklund, C., R. B. Marks, K. L. St andwood, and S. Wang, " I EEE St andard 802.16: A Technical Overview of t he WirelessMAN Air I nt erface for Broadband Wireless Access," I EEE Com m unicat ion Magazine, June 2002, pp. 98- 107. < Day Day Up >

< Day Day Up >

Part 4: Wireless Personal Area Networks Bluet oot h Specificat ion version 1.1 volum e 1, " Specificat ion of t he Bluet oot h Syst em : Core," February 2001. Bluet oot h Specificat ion version 1.1 volum e 2, " Specificat ion of t he Bluet oot h Syst em : Profiles," February 2001. I EEE 802.15 WG, I EEE P802.15.1/ D1.0.1, " Draft St andard for I nform at ion Technology— Telecom m unicat ions and inform at ion exchange bet ween syst em s—Local and m et ropolit an area net works—Specific requirem ent s—Part 15.1: Wireless Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions for Wireless Personal Area Net works ( WPANs) ," Sept em ber 2001. I EEE 802.15 WG, I EEE P802.15.3/ D10, " Draft St andard for Telecom m unicat ions and I nform at ion Exchange Bet ween Syst em s—LAN/ MAN Specific Requirem ent s—Part 15: Wireless Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions for High Rat e Wireless Personal Area Net works ( WPAN) ," June 2002. I EEE 802.15 WG, I EEE P802.15.4/ D13, " Draft St andard for Telecom m unicat ions and I nform at ion Exchange Bet ween Syst em s—LAN/ MAN Specific Requirem ent s—Part 15.4: Wireless Medium Access Cont rol ( MAC) and Physical Layer ( PHY) specificat ions for Low Rat e Wireless Personal Area Net works ( LR- WPANs) ," Decem ber 2001. < Day Day Up >

< Day Day Up >

Part 5: 2.5G and 3G Networks GPRS 3GPP TS 22.060, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Services and Syst em Aspect s; General Packet Radio Service ( GPRS) ; Service descript ion, St age 1 ( Release 5) ," March 2002. 3GPP TR 22.941, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Services and Syst em Aspect s; I P Based Mult im edia Services Fram ework, St age 0 ( Release 5) TS 22.228 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Services and Syst em Aspect s; Service requirem ent s for t he I P Mult im edia Core Net work Subsyst em , St age 1 ( Release 5) ," Novem ber 2001. 3GPP TS 43.051, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; Overall descript ion, St age 2 ( Release 5) ," April 2002. 3GPP TS 43.064, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GERAN; Digit al cellular t elecom m unicat ions syst em ( Phase 2+ ) ; General Packet Radio Service ( GPRS) ; Overall descript ion of t he GPRS radio int erface, St age 2 ( Release 5) ," April 2002. 3GPP TS 44.118, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM EDGE Radio Access Net work; Mobile radio int erface layer 3 specificat ion, Radio Resource Cont rol ( RRC) Prot ocol, I u Mode ( Release 5) ," April 2002. 3GPP TS 44.018, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; Mobile radio int erface layer 3 specificat ion; Radio Resource Cont rol Prot ocol ( Release 5) ," April 2002. 3GPP TS 44.060, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; General Packet Radio Service ( GPRS) ; Mobile St at ion ( MS) - Base St at ion Syst em ( BSS) int erface; Radio Link Cont rol/ Medium Access Cont rol ( RLC/ MAC) prot ocol ( Release 5) ," May 2002. 3GPP TS 44.160, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; Mobile St at ion ( MS) —Base St at ion Syst em ( BSS) int erface; Radio Link Cont rol/ Medium Access Cont rol ( RLC/ MAC) prot ocol; I u m ode ( Release 5) ," March 2002. 3GPP TS 45.005, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group GSM/ EDGE Radio Access Net work; Radio t ransm ission and recept ion ( Release 5) ," April 2002. UM TS 3GPP TS 23.002, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Services and Syst em s Aspect s; Net work archit ect ure ( Release 5) ," June 2002. 3GPP TS 25.211, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio

Access Net work; Physical channels and m apping of t ransport channels ont o physical channels ( FDD) ( Release 5) ," June 2002. 3GPP TS 25.301, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Radio I nt erface Prot ocol Archit ect ure ( Release 5) ," June 2002. 3GPP TS 25.302, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Services provided by t he physical layer ( Release 5) ," Decem ber 2002. 3GPP TS 25.308, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; High Speed Downlink Packet Access ( HSDPA) ; Overall descript ion, St age 2 ( Release 5) ," March 2002. 3GPP TS 25.321, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; MAC prot ocol specificat ion ( Release 5) ," June 2002. 3GPP TS 25.322, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Radio Link Cont rol ( RLC) prot ocol specificat ion ( Release 5) ," June 2002. 3GPP TS 25.323, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Packet Dat a Convergence Prot ocol ( PDCP) Specificat ion ( Release 5) ," June 2002. 3GPP TS 25.331, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; Radio Resource Cont rol ( RRC) ; Prot ocol Specificat ion ( Release 5) ," June 2002. 3GPP TS 25.401, " 3rd Generat ion Part nership Proj ect ; Technical Specificat ion Group Radio Access Net work; UTRAN Overall Descript ion ( Release 5) ," June 2002. cdm a 2 0 0 0 3GPP2 C.S0001- C, " I nt roduct ion t o cdm a2000 St andards for Spread Spect rum Syst em s Release C," May 2002. 3GPP2 C.S0002- C, " Physical Layer St andard for cdm a2000 Spread Spect rum Syst em s Release C," May 2002. 3GPP2 C.S0003- C, " Medium Access Cont rol ( MAC) St andard for cdm a2000 Spread Spect rum Syst em s Release C," May 2002. 3GPP2 C.S0004- C, " Signaling Link Access Cont rol ( LAC) St andard for cdm a2000 Spread Spect rum Syst em s Release C," May 2002. 3GPP2 C.S0005- C, " Upper Layer ( Layer 3) Signaling St andard for cdm a2000 Spread Spect rum Syst em s Release C," May 2002. Sa t e llit e Com m u n ica t ion ETSI EN 301 790, " European St andard ( Telecom m unicat ions series) , Digit al Video Broadcast ing

( DVB) ; I nt eract ion channel for sat ellit e dist ribut ion syst em s," Decem ber 2000. ETSI TR 101 790, " European St andard ( Telecom m unicat ions series) , Technical Report Digit al Video Broadcast ing ( DVB) ; I nt eract ion channel for Sat ellit e Dist ribut ion Syst em s; Guidelines for t he use of EN 301 790," Sept em ber 2001. ETSI TR 101 202, " European St andard ( Telecom m unicat ions series) , Digit al Video Broadcast ing ( DVB) ; I m plem ent at ion guidelines for Dat a Broadcast ing," February 1999. ETSI EN 301 192, " European St andard ( Telecom m unicat ions series) , Digit al Video Broadcast ing ( DVB) ; DVB specificat ion for dat a broadcast ing," January 2003. ETSI ETS 300 802, " European St andard ( Telecom m unicat ions series) , Digit al Video Broadcast ing ( DVB) ; Net work- independent prot ocols for DVB int eract ive services," Novem ber 1997. I SO/ I EC 13818- 1, " I nform at ion Technology—Generic Coding of Moving Pict ures and Associat ed Audio: Syst em s, Recom m endat ion H.220.0," Novem ber 1994. < Day Day Up >