Digital Storage in Consumer Electronics: The Essential Guide 0750684658, 9780750684651

Citation preview

Digital Storage in Consumer Electronics

This page intentionally left blank

Digital Storage in Consumer Electronics The Essential Guide

Thomas M. Coughlin

AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Newnes is an imprint of Elsevier

Newnes is an imprint of Elsevier 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA Linacre House, Jordan Hill, Oxford OX2 8DP, UK Copyright © 2008, Elsevier Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone: (+44) 1865 843830, fax: (+44) 1865 853333, E-mail: [email protected]. You may also complete your request online via the Elsevier homepage (http://elsevier.com), by selecting “Support & Contact” then “Copyright and Permission” and then “Obtaining Permissions.” Recognizing the importance of preserving what has been written, Elsevier prints its books on acid-free paper whenever possible. Library of Congress Cataloging-in-Publication Data Coughlin, Thomas M. Digital storage in consumer electronics : the essential guide / Thomas M. Coughlin. p. cm. Includes bibliographical references and index. ISBN-13: 978-0-7506-8465-1 (pbk. : alk. paper) 1. Computer storage devices. 2. Household electronics. 3. Digital electronics. I. Title. TK7895.M4C68 2008 004.5—dc22 2007038608 British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. ISBN: 978-0-7506-8465-1 For information on all Newnes publications visit our Web site at www.books.elsevier.com 08 09 10 11 12 13

10 9 8 7 6 5 4 3 2 1

Printed in the United States of America

Working together to grow libraries in developing countries www.elsevier.com | www.bookaid.org | www.sabre.org

This book is dedicated to my mother, Connie, who taught me to keep working until a project is done. It is also dedicated to my family: Fran, Will, and Ben, for putting up with my time away from them to work on this book.

This page intentionally left blank

Contents About the Author ........................................................................................ xi Acknowledgments ...................................................................................... xiii Chapter 1: The Consumer Electronics Storage Hierarchy ................................. 1 Objectives in This Chapter ........................................................................................ 1 1.1 Introduction ....................................................................................................... 1 1.2 Growth in Digital Content Drives Storage Growth ......................................... 2 1.3 Economics of Consumer Devices ..................................................................... 6 1.4 Rules for Design of Digital Storage in Consumer Electronics ........................ 9 1.5 Classification of Devices Using Storage in the Home ................................... 11 1.6 Consumer Electronics Storage Hierarchies .................................................... 12 1.7 Multiple Storage and Hybrid Storage Devices ............................................... 17 1.8 Chapter Summary ........................................................................................... 22

Chapter 2: Fundamentals of Hard Disk Drives ............................................. 25 Objectives in This Chapter ...................................................................................... 2.1 Basic Layout of a Hard Disk Drive ................................................................ 2.2 Hard Disk Magnetic Recording Basics .......................................................... 2.3 How Data Is Organized on a Hard Disk Drive .............................................. 2.4 Hard Disk Drive Performance and Reliability ............................................... 2.5 Hard Disk Drive Design for Mobile and Static Consumer Electronics Applications ............................................................................. 2.6 The Cost of Manufacturing a Hard Disk Drive ............................................. 2.7 Disk Drive External Interfaces ....................................................................... 2.8 Hard Disk Drive Technology Development ................................................... 2.9 Chapter Summary ...........................................................................................

25 25 29 32 34 36 39 40 44 51

Chapter 3: Fundamentals of Optical Storage ............................................... 53 Objectives in This Chapter ...................................................................................... 3.1 Optical Disc Technologies .............................................................................. 3.2 Basic Operation of an Optical Disc Drive ..................................................... 3.3 How Data Is Organized on an Optical Disc ...................................................

53 53 56 58

www.newnespress.com

viii

Contents 3.4 3.5 3.6 3.7 3.8

Optical Disc Form Factors .............................................................................. Optical Product Reliability ............................................................................. Holographic Recording ................................................................................... Optical Disc Storage Development ................................................................ Chapter Summary ...........................................................................................

60 61 62 65 70

Chapter 4: Fundamentals of Flash Memory and Other Solid State Memory Technologies ................................................................................. 73 Objectives in This Chapter ...................................................................................... 73 4.1 Development and History of Flash Memory .................................................. 74 4.2 Erasing, Writing, and Reading Flash Memory ............................................... 75 4.3 Difficulties that Cause Wear in Flash Memory .............................................. 77 4.4 Common Flash Memory Storage Technologies: NOR and NAND .............. 78 4.5 Bit Errors in NAND Flash .............................................................................. 83 4.6 Managing Wear in NAND and NOR ............................................................. 83 4.7 Bad Block Management .................................................................................. 85 4.8 Embedded Versus Removable NAND Flash ................................................. 85 4.9 Flash Memory File Systems ........................................................................... 86 4.10 Single Level Cell and Multilevel Cell Flash Memory ................................... 86 4.11 Another Approach to Multilevel Cells ........................................................... 88 4.12 Trade-offs with Multilevel Flash Memory ..................................................... 90 4.13 Types of Flash Memory Used in Consumer Electronics Devices ................. 91 4.14 Flash Memory Environmental Sensitivity ...................................................... 91 4.15 Using Memory Reliability Specifications to Estimate Product Lifetime ...... 92 4.16 Flash Memory Cell Lifetimes and Wear Leveling Algorithms ..................... 93 4.17 Predicting NAND Bit Errors Based upon Worst-Case Usage ....................... 95 4.18 Flash Memory Format Specifications and Characteristics ............................. 96 4.19 Flash Memory and Other Solid State Storage Technology Development ..... 98 4.20 Expected Change in Cost per Gigabyte of Flash Memory Formats ............ 101 4.21 Other Solid State Storage Technologies ....................................................... 102 4.22 Chapter Summary ......................................................................................... 103

Chapter 5: Storage in Home Consumer Electronics Devices ......................... 107 Objectives in This Chapter .................................................................................... 5.1 Introduction ................................................................................................... 5.2 Personal Video Recorders and Digital Video Recorders ............................. 5.3 Home Media Center and Home Network Storage ....................................... 5.4 Chapter Summary .........................................................................................

107 107 108 120 129

Chapter 6: Storage in Mobile Consumer Electronics Devices ........................ 131 Objectives in This Chapter .................................................................................... 131 6.1 Introduction ................................................................................................... 131

www.newnespress.com

Contents 6.2 6.3 6.4 6.5 6.6 6.7

Automobile Consumer Electronics Storage ................................................. Mobile Media Players ................................................................................... Cameras and Camcorders ............................................................................. Mobile Phones .............................................................................................. Other Consumer Devices .............................................................................. Chapter Summary .........................................................................................

ix 132 141 158 164 170 172

Chapter 7: Integration of Storage in Consumer Devices .............................. 175 Objectives in This Chapter .................................................................................... 7.1 Introduction ................................................................................................... 7.2 Storage Costs in Consumer Product Design ................................................ 7.3 Development of Common Consumer Functions .......................................... 7.4 Intelligence of Digital Storage in Consumer Electronics ............................ 7.5 Matching Storage to Different Applications ................................................ 7.6 The Convergence of Electronics—When the Storage Becomes the Device—Or Was It the Other Way Around? ........................................... 7.7 Road Maps for Consumer Electronics Application Integration in Storage Devices ........................................................................................ 7.8 Chapter Summary .........................................................................................

175 175 176 179 182 187 187 191 199

Chapter 8: Development of Home Network Storage and Home Storage Virtualization .............................................................................. 201 Objectives in This Chapter .................................................................................... 8.1 Introduction ................................................................................................... 8.2 What Drives Home Networking Trends? ..................................................... 8.3 Networking Options in the Home ................................................................ 8.4 Push Versus Pull Market for Home Networks ............................................. 8.5 Home Networks for Media Sharing ............................................................. 8.6 Home Networks for Personal Reference Data Backup ................................ 8.7 Projections for Home Network Storage ....................................................... 8.8 Design of Network Storage Devices ............................................................ 8.9 Advanced Home Storage Virtualization ....................................................... 8.10 Home Network Storage and Content Sharing within the Home ................. 8.11 Privacy, Content Protection, and Sharing in Home Network Storage ......... 8.12 Chapter Summary .........................................................................................

201 202 202 204 206 207 209 212 214 217 221 222 223

Chapter 9: The Future of Home Digital Storage ......................................... 225 Objectives in This Chapter .................................................................................... 9.1 Digital Storage Requirements for Home Data Sharing and Social Networking ................................................................................................ 9.2 Integrated Multiple Purpose Devices Versus Dedicated Devices ................ 9.3 Physical Content Distribution Versus Downloads and Streaming ..............

225 225 237 239

www.newnespress.com

x

Contents 9.4 9.5 9.6 9.7 9.8 9.9 9.10

Personal Memory Assistants ......................................................................... Digital Storage in Everything ....................................................................... Home Storage Utility—When All Storage Devices Are Coordinated ........ Digital Storage in Future Consumer Electronics ......................................... Projections for Storage Demands in New Applications .............................. Digital Storage as Our Cultural Heritage ..................................................... Chapter Summary .........................................................................................

240 242 243 247 249 254 256

Chapter 10: Standards for Consumer Electronics Storage ........................... 257 Objectives in This Chapter .................................................................................... 10.1 Digital Storage Standards ........................................................................... 10.2 Consumer Product Standards ...................................................................... 10.3 Home Networking Standards ...................................................................... 10.4 Needed Standards for Future Consumer Electronics Development ...........

257 257 264 266 273

Appendix A ............................................................................................. 277 Appendix B ............................................................................................. 279 Appendix C ............................................................................................. 281 Bibliography ............................................................................................ 285 Index ...................................................................................................... 289

www.newnespress.com

About the Author Thomas M. Coughlin is the Founder and President of Coughlin Associates. Tom has over 25 years of experience in the data storage industry as a working engineer and high level technical manager. In addition to regular technical and management consulting projects, he is the publisher of many reports covering technology and applications for digital storage devices and systems. He has many published papers, reports, and articles. He has six patents on magnetic recording and related technologies. Tom is the founder and organizer of the annual Storage Visions Conference, a partner event to the International CES. Tom is a senior member and 2007 Chairman of the Santa Clara Valley IEEE Section and was chairman of the Santa Clara Valley IEEE Consumer Electronics Society in 2006 and past chairman of the SCV IEEE Magnetics Society more than once. He is also a member of APS, AVS, IDEMA, SNIA, AAAS, TCG, and SMPTE. Tom has a BS in Physics, an MSEE from the University of Minnesota, and a PhD in Electrical Engineering from Shinshu University in Nagano, Japan. For more information on Coughlin Associates go to www.tomcoughlin.com. For more information on the Storage Visions Conference go to www.storagevisions.com.

www.newnespress.com

This page intentionally left blank

Acknowledgments An undertaking such as this book requires lots of support from one’s peers as well as one’s family. Many folks reviewed and suggested content for the various chapters. In Chapter 2 Dick Zech provided significant input and material for the optical drive section. Jim Handy provided excellent illustrative figures, descriptions and other material to describe the action of flash memory devices in Chapter 3. Rick Wietfeldt reviewed Chapters 5 and 6 and helped get permission to use the TI block diagrams. Bert Haskell also reviewed Chapters 5 and 6 and gave me guidance in determining costs for consumer products. Michael Willett reviewed the Trusted Computing Group material in Chapter 7. Tips and hints came from many sources. Overall reviewers of the book include Tom Clark and Brian Berg. Thank you for all your suggestions and input. Thanks to Don Stroud and Portelligent for teardown pictures and information on several consumer devices. Also thanks to Peek Inside and iFixit for teardown images. Thanks to iSuppli for optical disc and bill of materials information. Thanks to Pat Hanlon for use of illustrations and other material for the phases of storage integration into consumer products. All good ideas and suggestions included in these pages are inspired by these folks and many others. I take responsibility for all the mistakes and omissions.

www.newnespress.com

This page intentionally left blank

CHAPTE R 1

The Consumer Electronics Storage Hierarchy Objectives in This Chapter •

Develop an appreciation for the role of digital storage in the growth of consumer electronics—higher content resolution is a direct function of digital storage capacity available.

•

Understand the key role of product price in the growth of consumer products for most markets and how this is impacted by the cost of the digital storage used.

•

Get an initial exposure to the development of standard consumer device functions and how these could be integrated more tightly with the digital storage devices.

•

Present some rules for the design of digital storage into consumer devices in order to create more successful products.

•

Review the role of memory in processor execution as well as mass storage.

•

Develop the concepts of a digital storage hierarchy for mobile and static consumer applications.

•

Demonstrate the advantages of using multiple storage devices in consumer products creating hybrid storage products.

1.1 Introduction Memory is a key element in the design of modern digital electronics. Memory allows the retention of information that can be used by the electronic system. Memory is

www.newnespress.com

2

Chapter 1

information. This information may be computer files but it can also be digital photographs, home videos, movies, music, and other personal and commercial content. The content is what makes these devices useful, and digital storage is where the content resides. Electronic memory may be temporary as is the case for volatile memory, where the information it contains disappears when the power is turned off, or it may be long term memory, or nonvolatile memory. Memory may be fast or slow, low power consumption or high power consumption and it may be inexpensive per byte or expensive. It may be part of the microprocessor electronics, a peripheral chip, an internal mass storage unit, or an external digital storage device. As we shall see in this chapter all of these types of memory play a key role in the design of digital devices. Working together in the proper way, different memory devices can give optimal application performance for the best possible price to the consumer. We shall briefly look at expected growth of digital storage demand for modern consumer devices. Then we will examine the unique environments for digital storage in static and mobile consumer devices. We shall explore the economics of consumer electronics products and look at factors that make people buy these devices as well as different ways that these devices can be obtained. Finally, we shall develop a finer understanding of memory hierarchies for static and mobile consumer devices as well as look at how different memory can be combined to provide greater value than either alone. We shall also see how digital storage devices are dependent on each other and how growth in unit numbers and capacity of one storage device can help drive demand for another.

1.2 Growth in Digital Content Drives Storage Growth Demand for digital storage is driven by the growth in personal family content, by the growing number of entertainment devices in and around the home that require digital storage, by the increasing resolution of personal and commercial entertainment content, and by the growth in the number of channels that people can use to access all types of content. More consumers have digital still and video cameras, either as stand-alone devices or in convergence devices such as cell phones. Content creation devices such as digital still and video cameras are becoming a common standard application that can be built into

www.newnespress.com

The Consumer Electronics Storage Hierarchy

3

consumer devices. In addition to being ubiquitous, camera resolutions make digital storage demands higher with time. Families today have many ways to enjoy content. No longer must they be tied to a schedule for programs on the television. Content can be recorded from broadcast, cable, or satellite broadcasts on a digital video recorder to play back later. Audio or video content can be downloaded from the internet and listened to on a variety of static and mobile devices such as MP3 players and personal media players (PMPs). This content is likely to be stored on personal computers and possibly backed up to external storage devices. In an increasing number of homes, content may be stored in a network storage device and made available through a wired or wireless network to various static and mobile devices throughout the home. Even automobiles are increasingly using digital content for entertainment and navigation purposes. In the future, all of these mobile and static devices may be part of a far more comprehensive home storage network architecture. Content sharing and access is growing enormously. Content can be downloaded from the internet and watched now or later, it can be brought down to a cell phone through the mobile phone network, and it can be shared with everyone via the many growing social networking web sites. In all cases, the resolution of the content that people want is increasing. Content is compressed for two basic reasons. First, it is compressed to get through slow network connections and to speed up downloading. Second, it is compressed to conserve digital storage space required. Compression can be lossy or loss-less. Lossy compressed content cannot be reconstructed to its original resolution, whereas loss-less content can be reconstructed using decompression technology. With increasing internet connection speeds and the low cost of digital storage capacity today, we are likely to see changes in the resolution of content people may want. For instance, MP3 files are a lossy compressed format where up to 90 percent of the original content file is lost during compression. The compression is done with a very fine human hearing compression model so in a noisy environment, and with less than optimal acoustic equipment, you probably can’t tell that there is content missing. With faster internet speeds and low cost storage there is beginning to be a pronounced shift from MP3 to less compressed or even loss-less compressed music formats.

www.newnespress.com

4

Chapter 1 Table 1.1: Media Units Versus Storage Capacity for Various Resolution Photos, Music, and Video Files Object Size (MB)

1

2

4

150

667

4,160

Capacity (GB)

4 MP Photos

8 MP Photos

MP3 Songs

HiD Songs

VGA Video

DVD Video

5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 45,000 50,000 55,000 60,000 65,000 70,000 75,000 80,000 85,000 90,000 95,000 100,000 105,000 110,000 115,000 120,000 125,000

2,500 5,000 7,500 10,000 12,500 15,000 17,500 20,000 22,500 25,000 27,500 30,000 32,500 35,000 37,500 40,000 42,500 45,000 47,500 50,000 52,500 55,000 57,500 60,000 62,500

1,250 2,500 3,750 5,000 6,250 7,500 8,750 10,000 11,250 12,500 13,750 15,000 16,250 17,500 18,750 20,000 21,250 22,500 23,750 25,000 26,250 27,500 28,750 30,000 31,250

33 67 100 133 167 200 233 267 300 333 367 400 433 467 500 533 567 600 633 667 700 733 767 800 833

7 15 22 30 37 45 52 60 67 75 82 90 97 105 112 120 127 135 142 150 157 165 172 180 187

1 2 4 5 6 7 8 10 11 12 13 14 16 17 18 19 20 22 23 24 25 26 28 29 30

5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125

Table 1.1 shows some projections for the number of digital photographs, music files, and video files for different resolutions and device capacity that a user might require some time in the not-so-distant future. The gray areas for each media format show an estimate of the desired unit content ranges. These are based on an estimate of how much content an individual would like to have available on-demand from a local (e.g., portable) device. •

Up to 20,000 photo images.

•

Up to 10,000 songs.

•

Up to 100 movies.

www.newnespress.com

The Consumer Electronics Storage Hierarchy

5

Table 1.1 can give us an estimate of the acceptable size of storage for various applications at different resolutions. The following are some examples: •

A 4-megapixel photo viewer with 20,000 images needs 20 GB.

•

An 8-megapixel photo viewer with 20,000 images needs 40 GB.

•

A 10,000 song MP3 player needs 40 GB.

•

A 10,000 song high definition (HiD) player (like a compressed DVD audio) needs 1.5 TB.

•

A 100 movie player at VGA resolution needs 70 GB.

•

A 100 movie player at DVD resolution needs >400 GB.

•

A combination 20 K 4-Mpixel photo, 10 K MP3 song, 100 VGA movie player needs 130 GB.

•

A combination 20 K 8-Mpixel photo, 10 K HiD song, 100 DVD movie player needs 1.75 TB.

Based on Table 1.1 and estimates for the decline in storage costs over the next few years, the storage device for a 20 K 4-Mpixel photo, 10 K MP3 song, 100 VGA Personal Video Player (PVP; 130 GB) could be sold for less than $55 by 2010, enabling a consumer product with these characteristics to sell for under $200. With the additional integration of the storage device into the host product suggested in a later section of this book, the net cost would be even less, and a new finished product price of less than $150 should be possible. Overall storage required for commercial and personal content in the home should explode over the next few years. Personal content should grow even faster than commercial content. This will create a much greater demand for backup and archiving of personal family content. New technologies must be created to manage this content, organize it, and make sure that it is backed up. Backup could be in the home on external direct attached storage devices or increasingly in home network storage devices. Disaster recovery of personal content could be enabled by the growth of online providers of backup services, where content in the home is automatically backed up into a service on the internet.

www.newnespress.com

6

Chapter 1 + Volume

SSPL

- NPZ $99 $199 $299 Consumer Electronics Price

Figure 1.1: Sales Volume as a Function of Product Price.1

1.3 Economics of Consumer Devices We shall see that consumer devices are very sensitive to price. This creates a huge focus on reducing the overall product price in order to reach the most customers. However, there are lease model approaches for equipment that allow more expensive consumer equipment to be paid off over time. We shall also see how the retail channel puts pressure on initial product manufacturing cost, since there are usually several layers of markup that also have to be taken into account in arriving at the final product price to the consumer.

1.3.1 Consumer Product Price and Demand There are significant differences between the retail and service markets for consumer electronics devices. The retail market is primarily a push market, where retailers advertise products and offer various marketing approaches and discounts to persuade potential customers to make purchases. By contrast, the service market, typified by cell phone and cable companies, offers the consumer electronics hardware for free, at a discount, or on a low-cost lease basis in order to increase the services purchased as an ongoing subscription by the customer. The retail market for consumer electronics products is very price sensitive, and digital storage devices are often one of the more expensive components used in these products. Consumer products are often purchased with discretionary funds, so price is a very important factor in a purchase decision. Figure 1.1 represents the willingness of customers to purchase a consumer device as a function of product price. 1

Source: Cornice Inc.

www.newnespress.com

The Consumer Electronics Storage Hierarchy

7

Storage Unit Price Is Large Percentage of Total Bill of Materials (BOM) Cost

Hard Disk Drive Value

Hard Disk Drive Value

Hard Disk

$50.00

Consumer Electronics Device

$65.00

$91.13

30% Markup Distributor

Hard Disk Drive Value

Hard Disk Drive Value

$84.50

$109.85 MSRP $199.99 30% Markup

30% Markup Retailer

Consumer

(1.30 X 1.30 X 1.30) = 220% $199.99/220% = $90.90 BOM Cost $50.00/$90.90 = Hard Disk Drive is 55% of BOM

Figure 1.2: Consumer Storage Markup Through the Retail Distribution Chain. Note that SSPL means single spouse permission limit (what an adult could generally purchase without getting into trouble), while NPZ means no permission zone (what an individual could probably buy with no negative repercussions).2 Clearly the sales volume increases significantly as the sales price declines. The SSPL point is estimated to be between $100 and $200 retail while the NPZ is less than $99. Above $200 the purchase is likely to require at least some family discussion before a purchase is made in order to avoid financial problems or at least some level of heated discussion. These categories are suggestive since individual family buying approval patterns may differ.

1.3.2 Cost Markups in the Retail Sales Channel Figure 1.2 shows the markup in price from the initial product cost through the typical retail distribution chain. As can be seen in the figure, the amortized effective cost of a 2

From 2003 presentation by Cornice Inc.

www.newnespress.com

8

Chapter 1

digital storage device in the final sale of the consumer device can be more than twice that of the manufacturer’s storage product price. With every facility that a product goes through in reaching the consumer, there is an additional overhead and handling cost that must be assumed. These must be added to the price that the product sells for at the next level in the distribution chain and ultimately to the consumer. As shown in Figure 1.1, as the price goes up, the unit volume of sales generally goes down. The latter consequence is generally, but not always, true. If a product is seen as of much higher quality, or if it gives the user a greater feeling of status, then he or she may be willing to pay a higher price than simple market dynamics would indicate. This is the case, for example, with the Apple iPod products. These products are perceived as being “cool” and giving the user a higher status. As a consequence, the price of these products has not declined with time as they would in a normal market. In the long run, unless a company can maintain new technologies that continue to be perceived as cool, they will eventually lose this status and the price and profit margins tend to decline.

1.3.3 New Opportunities for Electronic Integration Technologies used in consumer electronics products come from many sources and have varied histories. GPS positioning technologies were originally developed by the military and only in the last 15 years have come into use by consumers. Today, GPS-based positioning capability has become one of those standard functions that are being built into many consumer electronics products. Often, consumer electronics product technologies are expensive to manufacture initially, and then go down in price as product yields increase and unit volumes go up. Optical disc media product introductions often follow this trend. Every new optical disc format from CD to blue laser discs (HD DVD and Blu-ray) started out costing over $1,000 per player, and decreased to less than $200 within about five years from initial commercial introduction. Consumer markets have market niches, one of which is high end consumers who, for example, put elaborate theaters in their homes. These early adoption consumers drive the initial market for products and often, but not always, determine whether a product will be successful in a broader market. There are many sources of consumer product ideas, but whatever the source, they often start out expensive initially and then decrease in price with time. With new technological development, such as less expensive memory, they may also develop

www.newnespress.com

The Consumer Electronics Storage Hierarchy

9

more capabilities with time for the same or lower prices. Digital storage is one of the key technologies for many of these applications and, as indicated, for many products it is one of the largest contributors to the cost of the device. The cost reduction of consumer products is often dependent upon lowering the cost of storage. Another important factor in the reduction of consumer electronics product cost is to integrate more of the application functions into fewer and fewer electronic components. This is enabled by smaller and smaller semiconductor line-widths. With smaller area required to make an electronic component, more functions can be integrated on a given chip. As discussed later in this book, electronic integration of standard consumer applications, and standard consumer oriented commands built into storage products, may allow for the achievement of even greater levels of product integration. Ultimately, digital storage and the electronics could essentially become one produced manufacturing unit. This integration of storage and applications, perhaps enabled by firmware (software programmable functions), could open a whole new era of manufacturing cost reductions. Tighter integration of the overall system architecture with shorter electronic lead lengths would allow faster, more reliable products. If multiple standard consumer electronics functions were built into integrated storage products, these devices could be very flexible, and could even be the basis of very powerful and useful convergence devices (multiple function products). If the firmware for accessing and customizing functions is easy to use and open standard-based, then the time to introduce new products, as well as the capability of changing products now in the field, could be greatly enhanced. In the consumer market, product introduction timing can make all the difference in product profitability. This could have a very great value for the industry.

1.4 Rules for Design of Digital Storage in Consumer Electronics Based upon observations of what works and what does not work in the design of digital storage in consumer electronics products, I would like to propose some rules for digital storage design. These rules, if followed, should help make such products more successful in the marketplace. These rules are: •

Use the most cost-effective storage component(s) that provides enough capacity for the application.

www.newnespress.com

10

Chapter 1 •

Never design a product that intentionally limits the available storage capacity to the customer—always allow a means of storage capacity expansion.

•

If appropriate, incorporate the advantages of multiple types of digital memory to achieve some of their individual advantages. Often a hybrid product using multiple types of storage is better for an application than a single storage device.

•

Use electronics and firmware to protect the customer’s content and battery life.

•

Make it easy to back up and copy data (storage is cheap, time is not!).

•

Give consumers a way to protect their personal content and privacy (encryption of data on the storage devices could help with this).

•

Make storage management and organization automatic, for instance, protect data and prevent replication of corrupt data.

•

Design the components including storage to provide the lowest total product cost (storage integration concepts could help here).

The basic concept behind these rules is to make the end product more attractive to the consumer in terms of price, features, and performance. Most of these are common sense observations. The basic idea is that for a product designer to be objective, his or her goal must be to provide the customer the best possible product at the best possible price. Providing the best performance at a certain cost may involve using hybrid devices that combine more than one type of storage to get better overall system advantages. In a world where digital content is requiring higher and higher resolution, it is good to give the customer access to, or at least an easy option to add, more storage capacity. This may mean larger onboard storage, but it could also involve the use of wired or wireless storage peripherals that the device can tap into to get reliable access to larger amounts of storage than the device itself can carry. User content is valuable. In an age where consumers often carry and use content creation devices, such as digital still and video cameras in mobile devices that could be subject to loss or theft, protecting this content from unauthorized access is important. Backing it up as soon as possible is a requirement for consumer devices. As consumermade content increases in volume, creating ways to automate the backup, archiving, and protection of this content becomes more critical. Likewise, creating ways to organize consumer content, including automatic metadata capture and creation, will make our lives easier as the sheer volume of personal content increases, but our time to manage

www.newnespress.com

The Consumer Electronics Storage Hierarchy

11

it, likely, becomes less rather than more. These functions need to be automated, probably in more advanced home storage network systems. As mentioned before, there will be great advantages available for net product cost, better product performance, and quicker time to market by leveraging electronic integration. This is made possible by denser transistor architectures in storage devices. The result is to build standard consumer electronics application commands into storage device firmware.

1.5 Classification of Devices Using Storage in the Home We classify five types of digital storage devices or devices that are heavily dependent upon digital storage used in and around the home, including mobile devices and automotive devices. These are: 1.

Active devices allow user interaction with other devices to exchange content files. These include computers, PDAs, and smart phones.

2.

Drone players retain content for play output after it has been downloaded from another local source, usually an active device.

3.

Direct attached external storage devices expand local storage of another device or allow for backup of content on the other device. These external direct attached storage devices may use USB, Firewire (IEEE 1394), eSATA, or other less common external interfaces.

4.

Network attached external storage devices provide a central content sharing device or centralized backup of content on other devices.

5.

Static or mobile personal content creation devices are digital still or video cameras or perhaps in the future a life-log device. Such personal content may then be saved on an active device, a direct attached storage device, or a networked storage device.

We will also define consumer devices as static, that is, not frequently moved and mobile, that is, often moved. Static consumer devices using digital storage include digital video recorders (DVRs), set-top boxes with DVR capability, home direct attached storage, and home network storage. Mobile consumer devices that use digital storage include many different types of portable music and video players, digital still and video cameras, and automobile entertainment and navigation systems.

www.newnespress.com

12

Chapter 1

1.6 Consumer Electronics Storage Hierarchies We will examine the uses of various digital memories in electronic devices to get an understanding of what sort of trade-offs designers have to make in their electronic designs. Then we will broaden our analysis to look at storage hierarchies for static and mobile consumer applications. We shall use these hierarchies to lay down some general guidelines on what storage to use for different applications, including how different storage devices can be used together to create hybrid storage devices that combine the good features of both types of storage.

1.6.1 Digital Memory for Device Process Execution Memory in digital devices can be used for the execution of microprocessor commands or for longer term storage. The three types of electronic memory used for process execution are shown in Figure 1.3. SRAM (static random access memory) is the fastest memory but also the most expensive. DRAM (dynamic random access memory) is not as fast as SRAM but less expensive and it is faster than NOR (the “neither or” logical operation) memory. NOR is the least expensive of these execution memories, but it is slower. SRAM keeps data in the main memory of a processing unit, without frequent refreshing, for as long as power is supplied to the circuit. SRAM is very fast, and its speed is closer to that of the CPU. However, SRAM is much more expensive than DRAM, and takes more space than DRAM for the same size of memory. An SRAM bit consists of four to six transistors, which is the reason for the bigger size compared to DRAM. The advantages of SRAM are speed and the lack of a need for constant data refresh. The disadvantages of SRAM are cost and size. SRAM is often used for CPU cache memory (both level 1 and level 2) because of its speed.

Execution Memory Spectrum Lowest S/GB

NOR

DRAM

SRAM

Lowest Latency

Lowest Solution Cost @ Targeted Power/Performance

Figure 1.3: Speed and Price Characteristics of Common Process Execution Memory Devices.

www.newnespress.com

The Consumer Electronics Storage Hierarchy

13

DRAM is the most often used RAM in computers. DRAM can hold its data if it is refreshed by a special logic circuit (the refresh circuit). The refreshing reads and rewrites the content of the DRAM memory frequently to prevent loss of the DRAM’s contents. Even though DRAM is slower than SRAM and requires the overhead of a refresh circuit, it is still much cheaper and takes about one quarter of the space of SRAM. DRAM is smaller and cheaper than SRAM because it uses one transistor per bit while SRAM uses four to six transistors per bit. In a DRAM, a capacitor holds an electrical charge if the cell contains a 1 or no charge if it contains a 0. The refresh circuitry reads the content of each cell (bit) and refreshes each one with an electrical charge before the content is lost. NOR is a type of flash memory. Flash memory is nonvolatile, which means that it does not need power to maintain the information stored in the chip. Flash memory offers slower read times compared to volatile DRAM memory. Reading from NOR flash is similar to reading from random access memory, provided the address and data bus are mapped correctly. Because of this, most microprocessors can use NOR flash memory as Execute In Place (XIP) memory, meaning that programs stored in NOR flash can be executed directly without the need to copy them into RAM. The capability of acting as a random access Read Only Memory (ROM) allows NOR to provide a low power, low cost option to DRAM or SRAM in consumer products. For mid-range and low end mobile phones NOR flash is quite common.

1.6.2 Mobile Device Consumer Electronics Storage Hierarchy The previous section discussed memory used in the execution of microprocessor operations. In this and the following section we will address longer term data retention with memory that acts as mass storage. Mass storage can include various semiconductor devices, as well as magnetic hard disk drives, tape drives, and optical memory. These mass storage products have characteristics that make one a better fit for a particular application than another. Optical drives and tape drives are often too large to fit into mobile devices (although there have been designs in the past with very small optical discs or tapes that could and did fit into mobile devices). In Table 1.2 we list some important characteristics of various mass storage devices as well as a relative ranking of these technologies by these criteria where 1 is best and 5 is worst. These various candidate mass storage devices have different attribute advantages. Also, it should be realized that the attributes listed in Table 1.2 depend on many variables not

www.newnespress.com

14

Chapter 1 Table 1.2: Ranking of Various Attributes for Common Consumer Electronics Mass Storage Technologies

Memory

Write Speed (Storage Unit)

Read Speed (Storage Unit)

Rewritability (Storage Unit)

Environmental Sensitivity (Shock, Vibration, Humidity)

Access Speed (Time to Get to Data Location)

Power Used

$/GB

Base Cost $

NOR Flash NAND Flash

3

1

5

1

1

1

5

1

2

2

4

2

2

2

4

2

Hard Disk Drive

1

3

1

5

3

3

3

4

Optical Disc

5

5

3

3

5

4

2

3

Magnetic Tape

4

4

2

4

4

5

1

5

Comments

Lowest Base Cost Low Base Cost, Fast Read Lowest $/GB for Internal Storage Device Lowest Cost for Content Distribution Lowest Media Cost for Rewriting

included in the table. For instance, hard disk drives use less power as the disk gets smaller, so a small hard disc drive may use much less power than 5.25-inch form factor optical disc drives or most tape drives. Many of these deficiencies for different types of storage technology can be reduced by the use of better electronics and firmware. For instance, the wear that can occur with rewriting of flash memory cells can be reduced with error correction software as well as wear reduction software that spreads writing around so device cells tend to wear more uniformly. Likewise, power saving firmware in a design using a small form factor hard disk drive can make it possible to get as long a battery life as a flash memory. Also, these are comparisons between the actual storage units within the storage device (such as cells in a flash memory). In practice, a hard disk drive could provide faster write speeds than a flash memory, depending upon the design of the flash memory unit. In any actual design there are trade-offs taken to create the most optimal solution to match that application’s needs. Computer scientists often refer to the characteristics of various memory devices as constituting a storage hierarchy. The concept of a storage hierarchy allows sorting

www.newnespress.com

The Consumer Electronics Storage Hierarchy

15

Cache Memory (SRAM)

Main Memory memory DRAM

Access Speed

Hard Disk Drive

Removable Memory (Optical Disc, Disk, USB)

Figure 1.4: Traditional Computer Storage Hierarchy. various memory products based on important attributes or characteristics for the applications for which they are to be used. Figure 1.4 gives a typical example of a computer storage hierarchy based upon data access speed. In this figure, cache memory is fast SRAM, while the main memory for the Central Processor Unit (CPU) or microprocessor is DRAM. Hard disk drives are slower than DRAM but faster than the removable devices such as optical discs or USB drives. Below is a list of important characteristics for consumer electronics storage devices that we will use to construct a mobile consumer electronics storage hierarchy. Mobile consumer electronics products include music and video players, cell phones, and many other products. •

Price.

•

Size.

•

Power.

•

Capacity.

•

Data rate.

•

Reliability and environment.

www.newnespress.com

16

Chapter 1 Table 1.3: Comparison of Mobile Storage Product Important Characteristics OEM Price Capacity Height Width Length Write Data Rate Read Data Rate Active Power Standby Power Temperature Range Operating Shock Non-Operating Shock Noise Washer Resistance

SFF HDD (1.8-inch)

MLC SD Flash

NOR Flash

$50–$90 40–80 GB 8.0 mm 54.0 mm 71 mm 16.6 MBps 16.6 MBps 1,100 mW 70 mW 0–70°C 500 G 1,500 G Some Low

$42–$85 4 GB–8 GB 1.4 mm 24 mm 32 mm 10 MBps 15 MBps 36 mW 0.2 mW −20–65°C >1,500 G >2,000 G None High

$20–$25 120 MB NA NA NA 0.145 MBps 200 MBps 90 mW 0.2 mW −25–85°C >1,500 G >2,000 G None High

The storage products that we wish to consider in this hierarchy are small form factor hard disk drives, NAND flash memory, and NOR flash memory. Table 1.3 compares these characteristics for these mobile storage devices at the time of writing. Storage capacities in this table are not representative of all products, and reflect the situation in 2007. Note that the last row, Washer Resistance, refers to the survivability of a removable storage device to a clothes washer cycle. Note that magnetic tape is used in many digital video cameras and could also be included in a mobile hierarchy. In Figure 1.5 we make an attempt to construct a mobile consumer electronics storage hierarchy. In this figure, we include NAND flash memory, small form factor hard disk drives, and optical memory. We include an optical disc since there are some optical technologies available that could provide a small form factor optical disc with many GB storage capacities. These could be used for content play or distribution. At the time of this writing there are no such products, so the properties of this optical disc are hypothetical.

1.6.3 Static Device Consumer Electronics Hierarchy For static or fixed consumer applications such as Digital Video Recorders (DVRs) or home media servers, the important considerations for the choice of digital storage

www.newnespress.com

The Consumer Electronics Storage Hierarchy Write Speed MB/s Write Life Multitasking

17

Cost $/GB

NAND Flash 10.6 $/GB 15 MB/s

Environmental Resistance Read Speed MB/s

Small SFF Form HDD Factor 1.12Disk $/GB Hard Drive 10.0 1.12MB/s $/GB 10.0 MB/s

Optical Disk ~0.2 $/GB

Figure 1.5: Mobile Consumer Electronics Storage Hierarchy.3 include storage capacity and price, read and write speed, ability to handle multiple content streams, vibration, acoustical noise, and long term reliability. The devices used for static consumer storage include hard disk drives, flash memory, and optical storage. There may also be some digital magnetic tape used in older consumer products. Figure 1.6 shows a digital storage hierarchy for static (e.g., stationary) consumer applications (after the method of S. R. Hetzler4). In this view, we show performance and general usage of digital storage devices for several consumer electronics applications that use different combinations of elements of the storage hierarchy.

1.7 Multiple Storage and Hybrid Storage Devices While single storage device consumer electronics products may give basic functionality, they may not provide all the features that customers want. This often leads consumer 3

4

S. R. Hetzler. The Evolving Storage Hierarchy. Presented at the INSIC Alternative Storage Technologies Symposium, Monterey, CA, 2005. Source: T. M. Coughlin. Development of Digital Storage for Consumer Electronics. Presented at ISCE Conference, June 2006.

www.newnespress.com

18

Chapter 1 DRAM HDD

Home Media Center with Tape

Optical Tape

DVR

DVR with Read/Write DVD

Decreasing Performance and Cost

Figure 1.6: Static Consumer Electronics Storage Hierarchy Showing a Performance or Data Access Speed Hierarchy. Arrows Represent the Storage Hierarchy Elements Used in Some Common Static Consumer Devices.

product designers to build into the device multiple storage devices or external connections to allow for digital storage expansion. This section will look at some of these multiple storage format consumer products as well as storage devices that are themselves a hybrid of two or more storage technologies.

1.7.1 Multiple Storage Format Consumer Devices There are many reasons why customers want multiple storage technologies in a single consumer product. For instance, when VHS tapes were phasing out and the red-laser DVD disks were starting to take over as the primary customer content distribution technology, many people had collections of VHS tapes that they still wanted to be able to play. This led to a thriving market for DVD players that also played VHS tapes. Digital video recorders are often shipped with only enough storage capacity for 30 hours or so of standard definition content (for HD content this would be much less). After recording for a few weeks, customers could easily fill the hard drive of a DVR, and be forced to erase older content to make way for new content. If the customer

www.newnespress.com

The Consumer Electronics Storage Hierarchy

19

didn’t want to erase the old content, he or she did not have many options. This problem has led to two solutions, one short term and the other long term. Both use multiple storage technologies to solve the problem of limited internal DVR storage capacity. A very popular category of DVR products, especially in Asia, has been a combination DVR and DVD recorder that can record content from the DVR onto optical discs so the content can be saved even if the hard disk drive in the DVR is erased. The other approach to expanding the storage on a DVR is to allow an external storage device to be attached to it that can expand the effective storage capacity of the DVR without replacing the internal hard disk drive. Many set-top and even stand-alone DVRs now come with an external SATA (eSATA) port on the back. eSATA allows data to be transported over the connector cables at rates up to 3 Gbps (375 MBps). Such very fast connections allow fast transport of even HD video content. This eSATA port allows the connection of an eSATA external storage device to the DVR. For instance a 1 TB (1,000 GB) hard drive external storage device with an eSATA interface could give significant additional storage capacity to a DVR with a 40 GB internal hard drive. Figure 1.7 shows an eSATA storage device attached to a set-top box. These are only a few examples of combining multiple storage technologies to create a more attractive product to the consumer. Such enhancements also have the beneficial effect for the consumer electronics company of allowing the sale of these combination products at a premium over and above the cost of adding the extra features. This helps create differentiation at first, and gives customers extra features that they really want. There are many other examples of combined storage products. Computers include a combination of many different storage technologies. They usually have a hard disk

Figure 1.7: eSATA Storage Expansion Box Attached to an eSATA Interface on a Digital Video Recorder Enabled Set-Top Box.

www.newnespress.com

20

Chapter 1

drive, at least one optical drive, multiple USB connections, and often places to directly plug in flash memory cards—usually to transfer photographs to the PC. This approach is a common one, allowing customers to make their choices of what they prefer and to allow for storage technology transitions. When USB drives first came out, computers for many years came with both USB connections as well as floppy disk drives. Today floppy disk drives are only available as USB external products to allow reading older content.

1.7.2 Hybrid Storage Devices In a way, a storage device that combines a solid state storage technology with another type of storage is a step toward application integration into storage product. For instance, hybrid hard disk drives (called ReadyDrive in the Windows Vista operating system) add a large NAND flash memory cache on the circuit board of a hard disk drive. This NAND cache can be used to save content that can be written to the disk drive later. This is the focus of the first generation of the hybrid hard disk drive. Later products using a large enough flash cache could store the working files being processed, with the hard drive serving as the backup for that data as well as the source for larger content that is not currently in the large NAND cache. Figure 1.8 shows first generation hybrid drive architecture. This technology is first being introduced into 2.5-inch hard disk drives for laptop computers. The disk drive controller is set up to handle the conventional volatile DRAM buffer found on hard disk drives, as well as a nonvolatile flash memory. Write data is cached into the nonvolatile flash memory and the disk drive is left in a low power mode until it is needed either to read data that it contains or when the nonvolatile flash cache becomes nearly full and its contents must be flushed to the hard disk drive to make room for more cached data. Data can be read from the disk via the DRAM, or directly from the nonvolatile cache memory. In addition to this caching function, some of the flash cache can be pinned, that is, kept in the cache. This would likely include some of the boot data for the operating system. Since the flash drive can be ready to transfer data faster than the hard disk drive, when the drive is powered up, the boot of the operating system can begin with the pinned data in the cache. A hybrid disk drive could leave the hard disk drive in a low power mode for a considerable time before the cache needs to be emptied to the hard drive. As a consequence, a hybrid hard disk drive can save a considerable amount of power when

www.newnespress.com

The Consumer Electronics Storage Hierarchy

Data Is Read into DRAM or into Flash Memory.

DRAM Buffer

HDD Flash

Surface

Controller

Write Data that Is Not “Pinned” Will Be Flushed to Disc.

21

Data Is Read from Disc (via DRAM) or Directly from Nonvolatile Memory.

Write Data Is Cached in the Nonvolatile Memory.

Figure 1.8: Hybrid Hard Disk Drive Architecture and Operation. compared to a non-hybrid hard drive. Even with a 50 percent decrease in the hard disk drive consumed power, the total power savings in the laptop may only be about 12 percent, because other components such as the display consume so much power. In addition, it can boot up a bit faster than the conventional laptop hard drive. In practice, the hybrid drive shaves a few seconds off a normal boot time. There are also Solid State Disk drives (SSDs) available for the mobile notebook market that use lower power than hard disk drives (although again the impact on total system power is minimized, since storage is a small part of total laptop power). SSDs also provide faster boot times. However, to have a storage capacity that approaches that of 2.5-inch hard disk drives (over 250 GB when this was written) the laptop computer would cost over $1,000 more. A hybrid hard drive also increases the shock and vibration endurance of the hard drive, since the drive disk is not spinning with the head loaded most of the time. Thus a hybrid product combines good features from multiple storage technologies and may represent an interesting architecture for consumer design, particularly mobile devices. I expect that next generation hybrid disk drives could be introduced for the high end video player market. These products could offer close to the current hard drive $/GB advantage but provide lower power usage, greater durability, and faster read performance.

www.newnespress.com

22

Chapter 1

Even SSDs may combine Single Level Cell (SLC) and Multilevel Cell (MLC) NAND flash memory together in a hybrid device. This hybrid device uses lower cost MLC storage for data that is frequently read, while the SLC memory is used for data that is often written. Since MLC flash sacrifices the number of possible erase cycles that it can go through to increase the flash cell storage capacity, this could give a larger storage capacity for a given price while helping with the long term reliability of the overall device. Hybridization of storage technologies can be a great way to leverage the advantages of the individual storage technologies. Successful hybridization requires good knowledge of the specifications, the performance of the individual storage technologies, and the possible interactions that the two technologies can have together. Hybrid storage is definitely a tool that consumer product designers want to keep handy in their design toolbox.

1.8 Chapter Summary •

Large digital storage capacities at an affordable price are key drivers of new consumer electronics products. With larger storage capacities, consumers can store and access ever larger collections of higher resolution content. The high end audio/video player of the future may easily require several hundred gigabytes of storage capacity.

•

The consumer electronics market is a very price sensitive market. Products with prices greater than about $200 generally sell lower volume than those that are less than $200, and the volume increases even further when the price drops below $100.

•

In many consumer products, the digital storage is a large part of the total system cost. The initial cost of the storage to the consumer electronics company is magnified by the overhead from the distribution chain to get to retail stores.

•

With greater levels of electronic integration possible, and with many consumer electronics functions becoming standardized, it should be possible to mate consumer applications and storage electronics to result in lower total product cost, greater reliability, and faster performance.

•

A number of common-sense rules have been developed for the design of digital storage in various products in the consumer industry.

www.newnespress.com

The Consumer Electronics Storage Hierarchy

23

•

We reviewed memory options and characteristics for cache and main memory of microprocessors.

•

Different digital storage devices have different advantages depending upon the application. Since static and mobile consumer devices have some common characteristics, we developed digital storage hierarchies that can be used to choose the proper digital storage for the application.

•

Different storage technologies can be combined in consumer products to meet technology transitions and to enhance a product so it is more appealing to the customer and to enhance the profitability of the consumer electronics company.

•

If one storage technology is directly combined into another storage product, we can produce a hybrid storage device. Hybrid hard disk drives combining flash memory as a cache on a hard disk drive are an example of this. Such hybrid products can provide some of the advantages of the individual storage technologies and should be carefully considered by consumer electronics design engineers.

www.newnespress.com

This page intentionally left blank

CHAPTE R 2

Fundamentals of Hard Disk Drives Objectives in This Chapter •

Learn about the history and uses of hard disk drives.

•

Describe the basic layout and operation of hard disk drives.

•

Clarify the organization of data on a hard disk drive.

•

Understand the self monitoring and reliability specifications for hard disk drives.

•

See how disk drives are designed for various consumer applications.

•

Enumerate the factors that determine the cost of a hard disk drive and its impact on CE product price.

•

Go over the developments of hard disk drive electrical interfaces.

•

Look at the future development of hard disk drive technology and the speed of areal density growth.

2.1 Basic Layout of a Hard Disk Drive Hard disk drives have a long history. Magnetic recording was first patented by the Danish inventor Vladimer Paulsen in 1898 using electromagnets and magnetic wires. Up to the 1950s (as well as later) engineers developed ways to use magnetic recording to record analog music and other recordings on various media including tapes and discs. In 1956 IBM introduced the first digital magnetic disk drive, the RAMAC (Random Access Method of Accounting and Control). The RAMAC had a total storage capacity of 5 million characters (somewhat less than 5 MB) on 50 24-inch disks in a device the size of

www.newnespress.com

26

Chapter 2

a large washing machine. It was leased annually to companies for several thousand dollars. Figure 2.1 shows the 1956 IBM RAMAC computer and storage system. Since 1956 hard disk drives have increased in capacity while decreasing in size and cost. In 2007 one terabyte of data could be purchased on a single 3.5-inch (95-mm) disk drive for less than $399. Hard disk drives have followed a relatively steady evolution of technologies that have allowed them to increase in storage areal density (the amount of digital storage capacity per usable surface area of the magnetic disk) at about 60 percent annually. For a given disk form factor this translates into a doubling of how much information can be stored on a single disk every 18 months. It should be noted that smaller disk drives are increasing in storage capacity faster than larger disk drives because the smaller disks have lower data rates making design at higher densities easier and also to give these products a continued edge over other storage technologies for mobile applications. Hard disk drives are used for a great many applications including high performance 15,000-rpm drives used in enterprise applications and high performance storage arrays. They are also used in near-line low cost high capacity storage arrays, desktop computers, and laptop computers. Of greatest interest to our readers is that they are used in an ever increasing number of static (mostly non-mobile) and mobile consumer applications. Hard disk drives are often the “mass storage” for the application that they are used in since they provide a large amount of non-volatile rewritable digital storage at a lower cost than all other storage devices except magnetic tape. Figure 2.2

RAMAC Hard Disk Drive System

Figure 2.1: The RAMAC Computer with the First Digital Hard Disk Drive System.

www.newnespress.com

Fundamentals of Hard Disk Drives 1,000,000

Mobile Consumer Electronics Desktop Enterprise ATA Enterprise

900,000 800,000 Units in Thousands

27

700,000 600,000 500,000 400,000 300,000 200,000 100,000 0 2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Figure 2.2: Hard Disk Drive Market Niche Projections to 2012.1 shows projections for hard disk drives for various applications through 2012 while Figure 2.3 presents projections for hard disk drive form factors through the same period. By 2013 or earlier over one billion hard disk drives should be in production per year. Consumer electronics and mobile computers will be the largest drivers of hard disk drive volumes and by 2010 smaller form factor hard disk drives (smaller than 3.5-inch hard disk drives) will be the majority of hard disk drives produced. A list of hard disk drive producers is given in Appendix C. Figure 2.4 shows a hard disk drive with the major components identified. The actuator motor moves the suspensions (at the ends of which the magnetic heads are attached) back and forth on the magnetic disk surfaces. The heads can then apply magnetic fields to write on the magnetic disk or read information back off the disk using magnetic read-back sensors. The sensors use nanotechnology devices such as giant magnetoresistive heads as well as tunneling magnetoresistive heads. The spindle motor rotates the disk beneath the head as the actuators move across the disk surface, allowing access to all of the user data on the hard disk drive. The heads have a surface texture that drives a very thin cushion (close to one millionth of an inch) of air under the head with the disk spinning. This air cushion supports the head close to but not contacting the disk surface as it reads and writes. The electronics on the 1

Source: Coughlin Associates, 2007.

www.newnespress.com

28

Chapter 2

1,000,000 900,000

Units in Thousands

800,000

1.3 inch or less 1.8 inch 2.5 inch 3.5 inch

700,000 600,000 500,000 400,000 300,000 200,000 100,000 0 2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Figure 2.3: Hard Disk Drive Form Factor Projections to 2012.2

Magnetic Disks

Drive Case

Actuator

Spindle Motor Suspension Assembly

Head

Figure 2.4: Major Components of a Hard Disk Drive, 1-Inch Disk Drive Shown.3

2 3

Source: Coughlin Associates, 2007. Image courtesy of Hutchinson Technology.

www.newnespress.com

Fundamentals of Hard Disk Drives

29

circuit board of a hard disk drive turns data into signals that are written on the hard disk drive or takes the electronic signals from the read sensors and turns it into digital data for the device using the hard disk drive.

2.2 Hard Disk Magnetic Recording Basics Hard disk drives record information on a magnetic layer laid down on the aluminum or glass disk substrate. A write element in the head projects a magnetic field into the magnetic recording layer that is high enough to change the prior magnetic state. As the digital information is written, the head field reverses in orientation, writing magnetic transitions between magnetized areas with opposite magnetic orientation. The oppositely oriented magnetic regions in the recording layer create fields above the disk surface that are detected by the read sensor in the head. As the recorded disk spins under the read head, the signals from the read sensors are decoded by the drive electronics to yield digital ones and zeros. A hard disk drive is generally designed to have a useful life of at least five years, although the actual drive specifications may differ depending upon the intended application and usage environment. From the invention of the digital hard disk drive in 1956 until 2005 the recorded magnetization in the magnetic recording layer was always oriented along the plane of the disk in commercial disk drives (this is called longitudinal recording). In 2005, Toshiba and Seagate introduced the first hard disk drives using perpendicular recording where the recorded magnetizations in the magnetic recording layer are normal to the disk surface. Figure 2.5 and Figure 2.6 are schematics of longitudinal and perpendicular magnetic recording, respectively. A hard disk drive has a pre-amplifier chip that is located on the head suspension or on the arm near to where the head suspension is attached. The pre-amplifier increases the voltage amplitude from the read sensor in the head, which is then transmitted by wires to the Printed Circuit Assembly (PCA) that is attached to the disk drive. Fifteen years ago a disk drive circuit board was a large complex board that occupied one entire side of a hard disk drive. Separate ICs were used for all the drive controls and the read/write channel. With the shrinkage of electronic circuit dimensions, it has become possible to incorporate more functions into a given IC. This has reduced the chip count on a hard

www.newnespress.com

30

Chapter 2

Giant Magnetoresistive Sensor

Track Width

N

S S

Shield Inductive Write Element

N N

Magnetization

S S

N N

S S

N N

Recording Medium S

Figure 2.5: Longitudinal Magnetic Recording Schematic.4

Recording Medium Magnetically Soft Underlayer

Figure 2.6: Perpendicular Magnetic Recording Schematic.5

4 5

Image courtesy of Ed Grochowski. Image courtesy of Ed Grochowski.

www.newnespress.com

Fundamentals of Hard Disk Drives

31

disk drive PCA and also reduced the number of discrete components that were once used to tune the circuits in a disk drive. Today System on Chip (SoC) designs for hard disk drives can incorporate many functions into one chip that formerly were done with separate ICs. As a consequence, modern hard disk drive printed circuit assemblies (PCAs) are much smaller, typically occupying only a fraction of one side of a 3.5-inch hard disk drive. Greater electronic integration has reduced the cost of hard disk drive PCAs, increased their reliability as well as durability, and allowed more functions to be incorporated in a hard disk drive than were possible in the past. Greater electronic integration has also made it possible to make hard disk drives as small as 1-inch and 0.85-inch form factors. Electronic integration in hard disk drives still continues. It is likely that hard disk drive boards will shrink further. Eventually we could have single chip designs with significant on-board intelligence. This could include built-in commands that can be used to create applications, lower production cost, and provide greater reliability. Greater intelligence built into hard disk drives is allowing them to function with lower power, generate less heat and noise, anticipate user actions, and provide advanced caching for writing and reading. Figure 2.7 shows the circuit board in a hard disk drive with key components labeled.

I/O Interface

System on Chip Read Channel

Motor Controller

Figure 2.7: Circuit Board of a Hard Disk Drive Showing Major Components.

www.newnespress.com

32

Chapter 2

2.3 How Data Is Organized on a Hard Disk Drive A hard disk drive is organized so that most of its surface is available for user data. There is also some surface area of the disks that store data used internally by the hard disk drive. The data on a hard disk drive consists of bits recorded along tracks on the disk surface. The total capacity of the hard disk drive is the product of the number of tracks times the average number of bits along those tracks. The tracks are recorded in concentric circles around the disk surface. These concentric tracks are broken into sectors of a fixed number of bits. Servo information is written on the tracks during the manufacture of the disk drive in recorded servo patterns at the beginning of sectors that give different signals as the head goes off-track in one direction or another. The servo signal is picked up by the head as the head actuator moves the head across the disk surface. The resulting signal is used to position the head over the track using a feedback circuit tied into the head actuator drive electronics. The sectors have drive accessible data recorded at their start, known as sector headers. The sector headers are used to identify the user data that follows and also provide parity data for error correction. Figure 2.8 illustrates the organization of data on a disk surface. The size of the disk sectors for many disk drives was fixed at 512 bytes for many years because of the requirements of popular computer operating systems, but changes in the

Sector Header

User Data

Figure 2.8: Data Is Organized on a Disk into Concentric Tracks that Are Broken up into Data Sectors.

www.newnespress.com

Fundamentals of Hard Disk Drives

33

Microsoft Vista operating system will enable using sector sizes as large as 4,096 bytes. The sector header information at the start of the sector must be read before the user data can be read. If the sectors are smaller, the sectors are located closer together, and the ratio of user data to drive data is lower, adding overhead to the percentage of the raw disk capacity that can be used for user data. By using larger sector sizes, higher capacity hard disk drives can be created, and sustained data rates off of the hard drive can be increased. The linear speed of disk tracks decreases going from the outer radius to the inner radius. In order to maximize storage capacity, disk drives use zoning to change the data rate to and from the disk during writing and reading as a function of the track radius. Zoning is done in order to maximize the recorded storage capacity on the disk. Modern disk drive circuits are also capable of optimizing the signal and recording density to accommodate specific capabilities of the actual heads and disks in the drive. Thus an optimum head and disk combination will record higher recording density than a lower performing head and disk combination. The advertised hard disk capacity is the formatted disk capacity that is accessible by the user. The actual raw storage capacity of a disk drive is higher since the drive can contain varying amounts of non-user-accessible data at the sector headers and in nonuser-accessible tracks. The hard disk drive also contains spare tracks and sectors that can be brought into use if an existing sector develops hard defects. Disk sectors containing hard defects that have been retired by the hard disk drive control circuitry are referred to as “g” list sectors. Most modern hard disk drives only allow access by users to the logical data structure on the disk and not the actual physical layout of the data on the disk. The mapping of logical data to the physical data is conducted internally by the hard disk drive and is transparent to the user. The development of virtualization of physical data to logical data has made it possible to create disk drives that are self-configurable and require much less work by the user in setting them up. This internal intelligence in hard drives was made possible due to the decreasing line-width of semiconductor processes. Narrower line-width circuitry allows more processing power and memory in modern hard disk drives (at the same time decreasing the actual size of the hard disk drive circuit board). Modern hard disk drives also contain capabilities for internal monitoring of important variables that can affect drive performance and reliability such as temperature and various types of defect data. Table 2.1 shows a list of common Self Monitoring and

www.newnespress.com

34

Chapter 2 Table 2.1: List of SMART Variables Monitored by Hard Disk Drives Attribute #

Attribute

Units

1 2 3 4 5 6 7 8 9 10 11 12 13 191 192 193 194 195 196 197 198

Raw Read Error Rate Throughput Performance Spin Up Time Start/Stop Count Reallocated Sectors Count Read Channel Margin Seek Error Rate Seek Time Performance Power-On Hours Spin Retry Count Recalibration Retries Device Power Cycle Count Soft Read Error Rate Gsense Error Rate Power on Retract Count Load/Unload Cycle Count Temperature HW ECC Recover Reallocation Event Count Current Pending Sector Count Uncorrectable Sector Count

199 200 201 220 221

UltaDMA CRC Error Count Write Error Rate Soft Read Error Rate Disk Shift G-Sense Error Rate

Read Errors — Sec or msec Spindle start stop counts Number of uncorrectable errors in sector Function not specified — Ave Performance Hours, Minutes, Seconds Only if First Spin Unsuccessful Only if First Attempt Unsuccessful Count of Power On/Off Cycles Program Read Errors — — Load/Unload or Landing Zone Cycles Degrees Centegrade — # Attempts to Transfer to Spare Sector # Sectors Waiting Remap Number of Uncorrectable Errors in Sector Errors in Data Transfer Errors While Writing Sector Program Read Errors Unknown Measurement of Shift Indication of Impact Loads

Reporting Technology (SMART) variables that hard disk drives support. Most hard disk drives also have built-in commands that will make the disk drive erase all the user data on the drive. This command, called secure erase, causes the drive to overwrite all the data on the drive so that it cannot be recovered. The secure erase still leaves the nonuser data intact so that the disk drive can be reused. Note that the SMART commands and monitored variables are part of the ATA specification.

2.4 Hard Disk Drive Performance and Reliability Hard disk drives access data by moving the heads across the tracks of the disk and by spinning the disk under the heads to find the location of a desired block of data. The disk drive keeps track of the location of data and how that data is connected via a host

www.newnespress.com

Fundamentals of Hard Disk Drives

35

computer-based file system. The data rate off hard disk drives depends upon the speed of rotation of the disk, the diameter of the disk, and the speed and accuracy of the head moving across the disk surface. For a 3.5-inch (95-mm) disk drive used in consumer applications such as digital video recorders, the sustained data rate of the hard disk drive is usually over 50 MBps. The sustained data rate is the rate for which the data from the disk drive can be continuously streamed off of the disk drive. Video content with standard definition (SD) has a data rate of about 6 MBps while high definition (HD) content data rate is >20 MBps. Since the data rate off of the hard disk drive is greater than that of individual SD or HD streams, the disk drive can support multiple streams. For instance, the hard disk drive can buffer content for one stream that is playing out. While that data plays out from the read buffer, the controller logic takes data being recorded from a write buffer and records it on the hard disk drive. In principle, with a 50 MBps sustained data rate, nine independent SD streams and two HD streams can be supported. In actual application, the overhead in changing from read to write will reduce the available streaming bandwidth and thus the number of streams that can be supported using a single disk drive. Hard disk drives are rated for their warranted lifetime and estimated time to failure. A typical expected useful life for a hard disk drive is about five years and warrantee periods vary from one to five years. The mean time to failure (or Mean Time Between Failures, MTBF) for hard disk drives used for consumer applications is generally 500,000 to 1,000,000 hours. Factors that can accelerate hard drive failure include: •

Start/Stop Cycles—number of disk motor On/Off cycles

•

Read Duty Cycle—percent of time drive reading during read stream

•

Write Duty Cycle—percent of time drive writing during write stream

•

Ambient Temperature

•

Other Environmental Conditions—such as vibration and shock

The number of start/stop cycles is influenced by the number of user On/Off cycles as well as power settings for a disk drive that may turn various elements On and Off as required. The number of read/write duty cycles is influenced by the size of read and write buffers, the required data rate to support continuous play, and the expected write/ read usage depending upon the application. As an example, a digital video recorder

www.newnespress.com

36

Chapter 2

(DVR; also known as a personal video recorder or PVR) typically does a lot of writing of new recorded content with somewhat fewer reads. The ambient temperature of the drive is controlled by heat dissipation in the host device by way of active cooling devices (fans) and passive cooling devices, the use and frequency of power saving modes, drive duty cycles, and the number of writes and reads. By monitoring the SMART log entries for internal drive temperature, usage may be adjusted to keep the internal temperature within acceptable limits. Shock is especially important for mobile applications. Vibration can be an issue for both mobile and static consumer applications. Moisture and humidity are generally bad for electronic products including hard disk drives. Temperature cycling in mobile and automotive applications can be an issue in drive performance and reliability since thermal expansion can cause off-track issues and condensation can cause reliability issues. It should be noted that more than half of all hard disk drive failures are due to circuit board or firmware failures while the balance of the failures are due to a variety of causes. This is why data on a disk drive can often be recovered using a different circuit board or with a firmware update. Note that the sensitivity to mechanical issues such as shock and vibration generally are less for smaller form factor hard drives since the inertia of the disks is reduced. Also, the total storage capacity, which is the product of the surface area and the areal storage density, is less for smaller form factor disks than for larger form factor disks. Heat generated by a hard disk drive and the power required to run it decreases as the disk radius gets smaller. In addition, smaller form factor hard drives are often quieter.

2.5 Hard Disk Drive Design for Mobile and Static Consumer Electronics Applications Mobile devices should be designed to be as small and lightweight as possible. A rule of thumb is to make the device with a specific gravity of 1 so that it seems like an extension of your own body. Thus mobile devices tend to go for smaller, lower mass storage products such as smaller form factor hard disk drives or flash memory. Today hard disk drives in the 2.5-inch form factor and smaller are used for mobile consumer applications. Drives in the 2.5-inch size are used in most laptop computers, although 1.8-inch form factor disk drives are used in ultra-portable laptops.

www.newnespress.com

Fundamentals of Hard Disk Drives

37

Disk drives of 2.5 inch and 1.8 inch are also being used in mobile external storage devices that can be connected to laptops and other products. Hard disk drives of 1.8 inch are used for higher capacity MP3 and video players as well as hard disk drive based video cameras. Hard disk drives that are 1 inch and smaller are used in some compact MP3 and video players as well as in external storage applications such as hard disk drive based compact flash form factor devices for very high resolution photographs. Seagate announced a technology called Digital Audio Video Experience (DAVE) that is an external mass storage device (the initial design demonstration at the 2007 CES used a 1-inch disk drive to provide up to 20 GB) in a credit card sized package with its own power supply as well as Bluetooth and Wi-Fi wireless connectivity. Such a device can be used to provide high capacity for local mobile devices such as mobile phones and cameras. Agere described a similar technology at the 2007 Storage Visions Conference, calling their technology Blue Onyx. This sort of device could be called a Personal Area Network Storage device (PANS). Small form factor hard disk drives used in mobile applications must be durable to avoid issues of shock and vibration that are common in mobile uses. They also must have power saving modes that turn the disk drive partially off when the drive is not being actively read from or written to. If the drive were left on with the disk spinning, it would rapidly use up the power stored in current battery power sources. Hard disk drive manufacturers have responded to the need for shock resistance in a number of ways. Parking the head off the disk when the disk is not spinning minimizes the risk of head crashes. They also implement design rules that leave sufficient sway space around the drive that helps to avoid shock. Using material that can absorb vibration as well as prevent vibration due to the mechanical motion of the disk drive increases the life of the drive. Manufacturers may also use sensors that detect motion of the hard disk drive and let the drive know it is falling. In that case the drive can be powered down and the heads retracted off the disk surface before the host device and the small hard disk drive reaches the ground. At the time of this writing 2.5-inch drives with 300 GB capacity and 1.8-inch drives with 160 GB capacity were coming to the market. Hard disk drives are also increasingly being used in automotive applications for navigation systems as well as entertainment. In these environments both vibration and the temperature range for operating and nonoperating conditions are extreme. Disk

www.newnespress.com

38

Chapter 2

drives designed for the automobile market are reduced in storage capacity from equivalent drives used in other applications. This is done by manufacturers as a strategy to reduce the track density (and to a lesser extent the linear density of the storage in the hard disk drive), which reduces the effects of track misregistration and run-out due to the changes in temperature and vibration. Thus digital storage capacity is traded off for a more robust hard disk drive. Static consumer applications include digital video recorders, desktop computers used at home, including home media center computers, and external storage boxes used for backup, or, if network capable, for sharing files. Static consumer applications are those in which the product containing the storage device is not moved often, hence it is static. Most static consumer applications use 3.5-inch hard disk drives since these storage products provide the highest storage capacity for the lowest price. At the time of this writing 3.5-inch hard disk drives were available with 1 TB of storage capacity on three disks. Much of the content in a home ends up ultimately on static storage devices. Even most mobile consumer products use a copy of MP3s or video files for which a master version is kept on a static storage product, often in a desktop or laptop computer. Because backup is becoming increasingly important in protecting irreplaceable personal content such as digital photographs or personal videos, there are also often multiple copies of content within a home (for instance the master copy, a copy on a mobile device, and a backup copy on an external storage device in case the drive with the master copy fails). A home disaster recovery market for backing up home content through the Internet to an external storage service (using hard disk drive arrays) is gradually developing as well. Thus if the house burns down and all copies of a family’s digital photographs and personal videos in the home are destroyed, the remote copies are still available and thus not lost. With the storage capacity of smaller form factor hard disk drives increasing with time (by at least 50 percent annually) these products may ultimately start to displace 3.5-inch hard disk drives for some applications, particularly where the size, weight, heat generation, and power consumption are important. A smaller form factor disk drive is better at all of these factors when compared to a 3.5-inch disk drive at the expense of lower storage capacity. Thus DVRs with 2.5-inch hard disk drives may start to show up for lower resolution content and even a desktop computer with 2.5-inch disk drives that will actually fit on your desk could become more common. Disk drives of 3.5-inch size will continue to be used for higher storage capacity products and also for external

www.newnespress.com

Fundamentals of Hard Disk Drives

39

Table 2.2: Typical Specifications for Hard Disk Drives Used in Common Consumer Products Application

Type

Capacity (GB)

Power (Peak, W)

Acoustical Noise (dB)

Maximum Shock (Operating, G)

Temperature Range (°C)

0.450

18

1,000

100

1.4

24

600

+5 to +55

65-mm HDD

40

6.5

24

800

−20 to +85

65-mm HDD 95-mm HDD

200

5.5

26

300

+5 to +55

1,000

6.2

29

55

+5 to +55

Portable 27-mm HDD 48-mm HDD

20

0 to +70

Car Home

storage devices for backup as well as additional digital storage capacity (e.g., to expand the storage available on a DVR so that older shows do not need to be erased to make room for new shows). Table 2.2 shows typical specifications for the common hard disk drive form factors used in consumer products.

2.6 The Cost of Manufacturing a Hard Disk Drive Hard disk drives have some standard components that are used no matter what the application or form factor. They all have a spindle motor to rotate the disk, they all have a head attached to a rotary actuator to place the head on the data that is read or written from the disk drive, and they all have at least one head and one disk. All have a mechanical housing to protect the hard disk drive components from the outside environment and they all have electronics on a circuit board that takes the analog voltages from or to the head to read or write information onto the magnetic disks, and that controls the motors and the interfaces conveying data back and forth from the hard disk drive to the host device. In addition, hard disk drives have firmware that is loaded to give the drive commands and basic functionality and they all go through a manufacturing, burn-in, and test process.

www.newnespress.com

40

Chapter 2

All of these steps add cost to a hard disk drive. For a 3.5-inch hard disk drive, the basic cost of building a minimum capacity hard disk drive (one disk and one head) is in the range of $30. To this cost of materials and direct manufacturing labor must be added overhead for the facilities and other corporate functions. A minimum break-even price of about $35 for a minimum capacity hard disk drive is likely. To make money, a drive company must sell this drive for some markup (say 20 percent) and then the Original Equipment Manufacturer (OEM), integrator, or the drive distributor or retailer can easily add another 20 percent or more on top of that. Thus the overall minimum price to the consumer of a minimum hard disk drive in the application is about $50.

2.7 Disk Drive External Interfaces The external interface to a hard disk drive connects the hard disk drive to the host device. The interface passes signals back and forth between the disk drive and the host. Associated with the physical connection or communication layer there are higher level functions that are associated with the interface, called higher level communication layers. For sophisticated communications systems such as Ethernet, there are several higher level layers that can provide very sophisticated service functions. In the case of today’s hard disk drives used in most consumer devices, the electrical interface does not support all the high level communication layers that true networking architectures such as Ethernet does, but they do support many types of commands and some simple local storage shared disk drive connections. The three major categories of hard disk drive interface in use today are the ATA interface (AT Attached—as in PC AT in the early 1980s), the Small Computer System Interface (SCSI) and the Fibre Channel interface. Disk drives and disk drive based storage systems can be used as Direct Attached Storage (DAS) where the storage device directly interfaces with the host or through a network where communication between the storage device and the client is through an intermediary network (which usually has several communication levels) and associated switching or routing hardware. SCSI and Fibre Channel interfaces are used on high performance hard disk drives used in higher-end direct connect and networked storage arrays. These arrays are collections of several disk drives in a box that together act as a sophisticated storage system usually using RAID or distributed file system architectures to create larger

www.newnespress.com

Fundamentals of Hard Disk Drives

41

storage capacity systems with higher I/O performance. Higher I/O performance enables faster data rates between the storage array and the host, while RAID provides the ability to deal with one, two, or sometimes up to three failed disk drives at one time without losing data. SCSI and Fibre Channel disk drives have higher data rates than ATA interface disk drives due to generally higher RPM (currently up to 15,000 RPM) and higher performance rotary positioner components used to place the recording and playback head above the disk surface. In order to achieve these high RPM values, these drives use several small form factor disks. This limits the overall storage capacity of SCSI and Fibre Channel disk drives (at the time of writing to at most 300 GB). Most SCSI and Fibre Channel disk drives are sold as direct connected storage devices or direct connected storage arrays for high performance workstations. Other SCSI and Fibre Channel disk drives are used in storage arrays that can be hooked up to a storage network allowing block level access to a large virtualized storage pool (where the complexity of the individual storage devices is hidden in the storage system). This block level access is similar to that within a typical hard disk drive so the networked storage array acts like a big hard disk drive to the file system located in front of it. Such a block-based storage system is called a Storage Area Network (SAN). Another type of storage system has one or more disk drives in a storage array or cluster behind a file system that is connected to a network such as an Ethernet network. Such a storage system providing files to a network is called Network Attached Storage (NAS). Today it is very common to have a single storage system that provides both block and file level access depending upon how the data is accessed. Such systems combine SAN and NAS functionality. The original Fibre Channel American National Standards Institute (ANSI) standard was established in 1994. It uses a serial interface to simplify the earlier parallel networking architectures that were used in high performance enterprise storage systems. Serial interfaces use two signal paths to communicate information to and from the client device and the storage device. The data in a serial interface is sent in consecutive bits packaged as sequential frames. Serial data is reconstructed at the ends of the signal paths using bus adapters. Parallel interfaces use many parallel signal paths where the data is sent down all the signal paths at one time and put together at the ends of the paths. Because of small

www.newnespress.com

42

Chapter 2

differences in effective path length between the various paths in a parallel connection, the data flowing in parallel down the signal paths gets out of phase. In order to reconstruct the data at the ends of the interface, these phase differences must be taken into account. Correcting the skew between the various signal paths in a parallel connector becomes harder as the data rate increases. This limits the data rate that parallel interfaces can carry. Serial interfaces do not have the skew issue of parallel path connectors, and with the advent of fast electronics for serializing and deserializing data, it is possible to carry very high data rates in a serial interface. Fibre Channel interfaces can be twisted pair copper or optical fiber and currently support up to four or more interfaces at 10 Gbps. Fibre channel interfaces are used in SAN networking as well as a native interface in some high-end (and very expensive) hard disk drives. The Fibre Channel specification uses the SCSI commands structure. The introduction of the Shugart Associates Systems Interface (SASI) in 1979 was an important milestone because it created an abstraction above the underlying disk drive by replacing physical addressing (cylinder, head, and sector) with logical addressing (logical block number). This relieved the host computer from the chore of bad block management, and allowed the physical characteristics of disk drives to change without impacting driver software in the host. This was possible due to the inclusion of a processor and software within the drive itself, something made possible by the everdecreasing cost of semiconductor technology. While SASI introduced this concept for disk drives, other types of peripherals were also beginning to include embedded processors. The need for an industry standard to support this kind of abstraction for multiple types of devices resulted in the formation of an ANSI standards committee in 1982. This committee finalized the SCSI standard in 1986. SCSI is a systems interface that supports devices, including disk, tape, CD-ROM, and printers, by way of common commands as well as commands that are tailored to the characteristics of each device type. In addition, this disparate set of device types could all coexist on the same physical bus. Additional sophistication, higher transfer speeds, and new device types were part of the SCSI-2 standard as finalized in 1994. Because of the advent of other storage physical interconnects that used serial technology, and because so much work had gone into creating the SCSI command set, SCSI-3 separated the logical command definitions from the physical interconnect definitions. This allowed the SCSI command set to be used with serial interfaces such as Fibre Channel.

www.newnespress.com

Fundamentals of Hard Disk Drives

43

With the official announcement of a Serial ATA interface (SATA—more on this below) in 2004 and with the limitations on speed of parallel interfaces, the ANSI standards committees that govern SCSI standards decided to launch their own serial interface, Serial Attached SCSI (SAS). Disk drives with SAS interfaces began to appear in 2006 and should become the common SCSI connectivity in the next few years due to the ease of cabling with serial connections and the faster data rates it enables. SCSI disk drives sell for a significant premium over ATA interface disk drives (usually two to three times higher) although the price of SCSI drives have been dropping to remain competitive with ATA interface disk drives used in many modern storage arrays. The current ATA interface began in 1986. It was the joint result of work by Control Data Corporation (Imprimis), Western Digital, and Compaq Computer. This interface has been called IDE by some companies and ATA by others. The first standards document on the ATA interface came out in 1989. Until 2005 all disk drives used various generations of the parallel ATA interface. Because of the skew issues discussed earlier, the parallel ATA interface limited the data rate of ATA. In the early 2000s, the ATA ANSI committee began serious work to come out with a serial interface standard capable of higher data rates. The initial serial ATA standard (SATA I) released in 2004 allowed up to 1.5 gigabit (or billion bits) per second (Gbps) or 187 megabytes per second (MBps) data rates. The 2.0 specification now allows up to 3.0 Gbps (375 MBps) data rates, making the SATA interface one of the fastest available, particularly to small system developers. The SAS interface was designed to be compatible with SATA for a properly designed common communication backplane. Thus higher performance, higher priced, but lower capacity SAS drives could be combined with higher capacity, lower priced, and lower performance SATA drives in a common storage system. Note that both SAS and SATA currently support up to 3 Gbps (or 375 MBps) data rates. Previously, parallel ribbon connectors were used for parallel ATA connections, but that has been updated to the newer SATA connectors. The serial interface connector is much easier to use than the parallel connector and also more reliable since it was all too easy to plug the parallel connectors in slightly askew so all the parallel connections were not made. The serial connector is also lighter and more flexible, making it easier to work with.

www.newnespress.com

44

Chapter 2

ATA interface disk drives are the most numerous disk drives produced in the world. They are used in desktop, laptop, and most consumer electronics applications. ATA interface disk drives used in low cost storage arrays are also the fastest growing portion of the enterprise storage market. Thus ATA disk drives enjoy advantages due to mass production and thus lowering of the cost of components and wide availability of suppliers. In addition to simplification of the interface with serial connectors, SATA also includes more sophisticated commands than the older ATA storage standards. There are many new commands in the SATA specification that are focused on multiple device connections, external or eSATA, and particular functions associated with common consumer functions such as support for multiple streams of content (i.e., for parallel write and read on a digital video recorder). SATA disk drives offer a low cost high performance interface for consumer applications that need the storage that a hard disk drive can provide. The CE-ATA specification was developed in 2005. It was designed to standardize connections for small form factor hard disk drives in mobile devices, such as 1-inch hard disk drives. By standardizing the connection for these drives, the design of hard disk drives to be used in consumer electronics devices is easier, the commands needed for mobile consumer electronics devices can be simplified, and commodity connectors can be used. In particular, the CE-ATA specification calls for using the MMC connector on the host device with a matching flex cable coming from the CE-ATA hard disk drive. Figure 2.9 shows a prototype CE-ATA interface with a hard disk drive. Table 2.3 shows the pin-out for the flex connector used in a 4X CE-ATA connector.

2.8 Hard Disk Drive Technology Development The storage capacity of a hard disk drive is the product of the total disk surface area available for data on the drive times the average areal density of the information storage on the disks minus any loss of user storage capacity due to formatting of the hard disk drive and other non-user-accessible data. The growth in the storage capacity of hard disk drives has been enabled primarily by the growth in the storage areal density. The areal density of information storage on a disk drive is the product of the average linear density that information can be recorded on a disk track and the average density of the data tracks. Increases in areal density are dependent upon scientific breakthroughs and

www.newnespress.com

Fundamentals of Hard Disk Drives

45

Figure 2.9: Prototype CE-ATA Interface Hard Disk Drive. Table 2.3: Pin Assignments for 4X Data Signal Lines in Flex Connector Pin # 1 2 3 4 5 6 7 8 9 10 11 12

Signal Name VSS DAT2 DAT3 Supply Voltage CDM Interface Voltage CLK VSS DAT0 DAT1 VSS Reserved

engineering integration of technologies that allow higher linear densities and track densities on the hard disk drives. Through the years, progress in hard disk drive technology has been measured through several parameters such as data rates, hard disk drive reliability, shock resistance, and other characteristics. The most significant characteristic that has driven the growth of

www.newnespress.com

46

Chapter 2

hard disk drive use in so many new applications, however, is the amount of data that can be packed into a hard disk drive with a finite number of heads and disks. By increasing the storage capacity of a hard disk drive using the same number of components, it has been possible to create a cost per gigabyte of storage that few other storage technologies can match. Some storage technologies such as magnetic tape and some optical recording devices have a low cost per gigabyte too, but they do not have the combination of attractive performance and convenience characteristics that have favored hard disk drives. At the time of writing, areal densities of shipping disk products exceeded 31.1 Gb/cm2 (200 Gb/in2) with fully characterized error rate areal density demonstrations exceeding 66.7 Gb/cm2 (430 Gb/in2). Recent technical advances suggest storage areal densities over the next decade will surpass 155 Gb/cm2 (1,000 Gb/in2). Other critical enablers for mass storage’s continued advances in capacity and performance are perpendicular recording, Heat Assisted Magnetic Recording (HAMR), availability of MicroElectromechanical Systems (MEMS) technology, smaller disk drive form factors, and a broad set of new consumer and appliance applications for mass storage data products. The year 2006 marked the second year of perpendicular recording drive shipments with the technology being geared up for all applications and form factors starting in 2007. Figure 2.10 shows a comparison of the actual quarter-by-quarter logarithmic areal density product announcement curve versus several Cumulative Annual Capacity Growth Rate (CAGR) curves starting from calendar first quarter 2000 (Q1 2000). The CAGR curves show the expected annual increase in areal density if the same rate of increase happens annually. As can be seen from these figures, it is possible to get a partial fit of the actual announcement data to one or another CAGR curves, but the fit is never very good and does not represent the rich complexity of actual technology development. Following are some observations on the introductions of new storage technologies based upon this example of hard disk drives: •

Technology introductions occur erratically in a monotonically increasing fashion, driven by the pace of technical discovery as well as introductions timed for maximum advantage to the introducing company.

•

Once a technology is introduced, it must go through a learning cycle until yield and performance issues are resolved, and then it follows a rapid adoption that displaces other technologies.

www.newnespress.com

Fundamentals of Hard Disk Drives

47

Gbits per Square Inch

1,000 100% CAGR 80% CAGR 60% CAGR 40% CAGR 20% CAGR HDD Product

100

Q 1 Q 20 2 00 Q 200 3 0 Q 20 4 0 Q 20 0 1 00 Q 20 2 01 Q 200 3 1 Q 20 4 0 Q 20 1 1 0 Q 20 1 2 0 Q 20 2 3 02 Q 200 4 2 Q 20 1 0 Q 20 2 2 0 Q 20 3 3 03 Q 200 4 3 Q 20 1 0 Q 20 3 2 0 Q 20 4 3 04 Q 200 4 4 Q 20 1 0 Q 20 4 2 0 Q 20 5 3 0 Q 20 5 4 05 Q 200 1 5 Q 20 2 0 Q 20 6 3 0 Q 20 6 4 06 Q 200 1 6 Q 20 2 07 20 07

10

Figure 2.10: Comparison of Product Announcement Trends Versus Areal Density Annual Growth Rates (Cumulative Annual Capacity Growth Rate, CAGR). •

There may be more than one approach that creates at least a short term solution to a given technological problem.

Figure 2.11 shows broad expectations of the generations of development of magnetic recording technology over the next few years indicating major developments in production heads and media used in disk drives. Perpendicular Magnetic Recording (PMR) extends the thermal stability of magnetic recording. In this list TMR is Tunneling MagnetoResistive reading and TAR is Thermally Assisted Recording (another term for HAMR). Patterned media is a way to create discrete regions of magnetic media on the disk where the data is stored. •

Generation 1: Initial product introductions by Toshiba and then Seagate

•

Generation 2: Current product generation—incremental improvements in PMR, greater use of TMR readers (products by all drive manufacturers)

•

Generation 3: Further improvements including discrete tracks and/or dual stage actuator

www.newnespress.com

48

Chapter 2

Areal Density in Gigabit per Square Inch

10,000.00 Longitudinal Magnetic Recording (LMR)

Generation 1 Generation 2 PMR PMR

Generation 3 PMR

Generation 4 PMR

1,000.00

100.00

20 10

20 09

Q 1

20 1 Q

Q 1

08

07 20

06 Q

1

20

05 1 Q

1

20

04 20 Q

1 Q

1

20

03

02 20 Q

01 1

20 Q

1 Q

Q

1

20

00

10.00

Figure 2.11: Projected Areal Densities and Recording Technologies in Production Hard Disk Drives. •

Generation 4: Advanced perpendicular recording media, TAR/HAMR or patterned media with more advanced read sensors, dual stage actuators

With the projected growth of product areal density by 2010 we could see single-disk hard drives with the following capacities for various form factors: •

100+ GB 1-inch disk drives.

•

300+ GB 1.8-inch disk drives.

•

500+ GB 2.5-inch drives.

•

1+ TB 3.5-inch drives.

Hard disk drive prices declined rapidly through the 1990s, forcing large scale consolidation in the industry but also opening many markets to the use of hard disk drives, such as consumer electronics. However, since 2004, the average price of hard disk drives has declined more slowly. With the average storage capacity increasing and with a five percent or so average price decline, we could see 3.5-inch disk drives with prices per gigabyte of $0.06 by 2010. No other storage technology except magnetic tape can claim such economical digital storage.

www.newnespress.com

Fundamentals of Hard Disk Drives

49

At the same time that areal density has increased so dramatically (say 6,000 times over the last 15 years), the time it takes to get to data on the disk drive has increased. The time to data on a hard disk drive is a function of the linear density of data on the hard disk drive tracks, the rotational speed (RPM) of the disk, and the layout of data on the hard disk drive. Hard disk drives are also increasing in reliability due to several factors. As the average form factor decreases, the mass of the hard disk drive decreases and inertial shock is less significant. Smaller drives also are less sensitive to and less likely to induce shock. In fact, the nonoperating shock of a 1-inch hard drive is about the same as a flash memory. In addition, smaller drives run cooler and use less power. When power-saving modes are used, the drives are not actively spinning and are thus very shock resistant. Even larger hard disk drives are using more intelligent electronics and control commands enabled by more sophisticated circuits on these devices. The inclusion of shock sensors and accelerometers are also enabling disk drives to detect when they are dropped so if the drive is in operation the heads can be taken off of the disks to prevent head slap and loss of data. All of these factors as well as better understanding of how to build a hard disk drive into a consumer electronics device are making hard disk drives more reliable. In addition to wanting lower storage costs, users of storage would like to get to their data faster as well as have lots of it stored economically and this will lead to some changes in hard disk drives. The introduction of Vista will bring some new supported options related to hard disk drives that may also help. Vista supports new write cache and read cache options that could improve the user storage experience. ReadyCache is a feature in Vista that allows plugging a USB flash device into a computer that can act as a read cache to hold frequently used data off of the hard disk drive and provide it faster than a hard disk could. Vista also supports ReadyDrive, which is a hard disk drive with a flash write cache inside the drive that holds information to be written to the hard disk drive (this is primarily for power savings in mobile products), thus increasing battery life. In addition, some of the operating system boot information can be stored on the flash memory and used to boot the computer faster than would be the case with the hard disk drive alone. Such hybrid hard disk drives as ReadyDrive may have applications in consumer electronics as well where the flash can serve as write or read cache to conserve power. Intel has a competing technology that they call Turbo Memory or Robson that puts the flash write cache on the system motherboard.

www.newnespress.com

50

Chapter 2

Another feature supported in Vista is an increase in the number of sectors of data on the hard disk drive from 512 bytes up to 4,000 bytes. The sectors break up the data on the tracks of the hard disk drive into chucks equal in size to the specified sector size. The sector size in hard disk drives has been limited to 512 bytes for many years for windows-based computers. Increasing the sector size in hard disk drives will allow higher storage capacities since the overhead for the sector header data will be less. This could also enable faster data access since the drive would not have to read as much sector data and thus could reach user data more rapidly. Larger sectors make a lot of sense for consumer applications where the data may be streamed. If the data is fragmented across multiple sectors, however, hybrid drives and larger sectors may not have nearly the same performance improvements. This brings up an important point in the use of digital storage in consumer devices. As the amount of digital data on consumer products increases and as this data is written and erased, the data on the storage device can become fragmented across the disk. This can increase the time to access the data and may make power-saving schemes ineffective. For this reason consumer products must include some internal storage intelligence and the capability to run management functions such as disk defragmentation in the background to preserve the performance and power usage of the consumer device over time. In general, as the electronics involved in building a hard disk drive become more sophisticated and free up more real estate on the hard disk drive circuit board, additional intelligence can be built into these storage devices. Some important steps in this development are SMART capabilities, consumer electronics commands in the ATA specification, and hybrid drives. If the trend to greater intelligence in hard disk drives continues we may see hard disk drives that contain some file system capability allowing them to know their own contents. Capabilities such as onboard drive searches could be done with such intelligent disk drives. If there are several intelligent disk drives then such searches could be done in tandem making a very fast search process possible. Internal file systems or even object storage on the hard drives could also allow more intelligent management of the disk drive and the data that it contains. More sophisticated communications capability could also be built into hard disk drives giving them new wired and wireless communications capability. Many consumer functions could even be designed into hard disk drives themselves making hard disk drive manufacturers Original Design Manufacturers (ODM) for consumer electronics companies. This could save money by eliminating a consumer electronics product circuit board and a separate

www.newnespress.com

Fundamentals of Hard Disk Drives

51

integration of consumer electronics and storage electronics and commands. If done with open standards this could also result in faster consumer electronics product designs and time to market and also better overall performance and CE consumer electronics product reliability. The development of thinking storage devices that are intelligent nodes on a home and mobile network is not far off. This concept will be explored further in chapter 8.

2.9 Chapter Summary •

Hard disk drives have been a key enabling digital storage technology for over 50 years. Hard disk drives make modern digital computers possible. Today hundreds of millions of these useful devices are manufactured per year for applications ranging from enterprise computing, desktop and laptop computers to static and mobile consumer applications.

•

Hard disk drives have data organized in sectors with the magnetic heads swinging across the disk to read and write information while the disk rotates beneath it. The disks are coated with magnetic material and the heads write and read information on the disks with magnetic fields.

•

Disk drives are designed to last several years when used according to their specifications. Different disk drives are designed with different specifications to match the needs of their intended application. Static applications such as digital video recorders use larger 3.5-inch disk drives while mobile applications use smaller form factor disk drives such as 1.8-inch disk drives.

•

Disk drives have developed electrical interfaces that provide very fast data transfer speeds. For SATA drives used in many consumer applications data can be transferred at 3 Gbps data rates today. Small portable devices may be built with the simpler CE-ATA specification.

•

The orientation of the magnetic material in the disk media has changed from in the plane of the disk (longitudinal) to out of the plane of the disk (perpendicular) in order to achieve higher magnetic recording densities. Future magnetic recording developments are expected to maintain at least a 50 percent annual increase in the areal density of magnetic recording. Magnetic recording areal density may be able to increase until it is several terabits per square inch. The highest shipping areal density at the time of writing is over 200 Gb/in2. Over 200 times current capacities of the various form factors may be possible using magnetic recording technology.

www.newnespress.com

This page intentionally left blank

CHAPTE R 3

Fundamentals of Optical Storage Objectives in This Chapter •

Go through the optical storage products now available on the market.

•

Understand the basic operation of optical drives and discs.

•

Teach the differences between CD, DVD, and blue laser optical storage systems and why the blue laser products can store more digital content.

•

Provide the quick overview of how data is organized on an optical disc.

•

Review the history of optical disc form factors.

•

Guide the user through the issues and expectations of optical disc reliability and longevity.

•

Learn about holographic recording technology and other advanced optical disc technologies.

•

Gain insight on what could drive the use of very large capacity optical disc technology.

3.1 Optical Disc Technologies Optical discs serve a variety of uses in consumer and computer applications. They are the most widespread medium for physical content distribution (CDs for music and DVDs for video). They are used in computers for backup and archiving of content as well as file sharing. Optical discs are used to record television programs that a user

www.newnespress.com

54

Chapter 3

wants to see later. They also allow users to burn discs containing personal videos to share with friends and family. Optical storage over the past 35 years has provided a multiplicity of storage solutions. The range spans the first analog video disc and 12-inch Write Once (WO) systems in the 1970s to today’s blue laser DVD drives and media. Optical storage is designed for specific consumer applications (primarily, digital audio and video for Read Only (RO) and rewritable CD and DVD media). Strict media standards (specifications) permit specific applications to be implemented by means of signal processing, logical and applications level software, and packaging. For example, DVD-Video is a consumer electronics application of DVD-ROM (a computer storage technology), not a separate format. A CD digital analog disc replicated in 1982 can still be played today, more than 20 years later. The same will likely be true for SDTV (standard definition television) DVD discs and successors. Table 3.1 summarizes the many types of optical disc storage shipping today in terms of primary

Table 3.1: Shipping Products (Examples of 2006 Optical Disc Storage Products) Product Name

Form Factor

Disc Type/ Diameter (mm)

Capacity (GB)

Write/ Read Options

In Production

Market Segments

CD-DA

5.25-inch HH 5.25-inch HH 5.25-inch HH 5.25-inch HH

Replicated/ 120 mm Replicated/ 120/80 mm Replicated/ 120 mm Dye Layer or Phase Change/ 120/80 mm Replicated/ 120/80 mm Dye Layer and Replicated/ 120 mm Replicated/ 120 mm Replicated/ 120 mm Replicated/ 120 mm Dye Layer/ 120/80 mm

0.65–0.80

RO

Yes

1

0.65–0.80

RO

Yes

2, 3, 4

0.65–0.80

RO

Yes

1

0.65–0.80

RO/WO/ RW

Yes

1, 2, 3

4.7–17

RO

Yes

2, 3, 4

0.65–0.80 4.7–17

WO/RO

Yes

2

4.7–17

RO

Yes

1

4.7–17

RO

Yes

1

0.65/4.7

RO

Yes

1

4.7–9.4

RO/WO

Yes

1, 2, 3

CD-ROM VCD/ SuperVCD CD-R/ CD-RW DVD-ROM DVD Combo (CD-RW + DVD-ROM) DVD-Video DVD-Audio SACD (dual layer) DVD+/-R

5.25-inch HH 5.25-inch HH slim-line 5.25-inch HH 5.25-inch HH 5.25-inch HH 5.25-inch HH

www.newnespress.com

Fundamentals of Optical Storage

55

Table 3.1: Continued Product Name

Form Factor

Disc Type/ Diameter (mm)

Capacity (GB)

Write/ Read Options

In Production

Market Segments

DVD+/-R DL

5.25-inch HH 5.25-inch HH 5.25-inch HH 5.25-inch HH 5.25-inch HH

Dye Layer/ 120 mm Phase Change/ 120 mm Phase Change/ 120/80 mm Phase Change/ 120 mm Phase Change/ 120 mm

8.5

RO/WO

Yes

1, 2, 3

4.7/9.4

RO/WO/ RW RO/WO/ RW RO/WO/ RW RO/WO/ RW

Yes

1, 2, 3

Yes

1, 2, 3

Yes

1, 2

Yes

1, 2

5.25-inch HH

Phase Change/ 120 mm

23.3

WO/RW

Yes

3

5.25-inch HH

Replicated/ Phase Change 120 mm Phase Change/ 120 mm

25/50 (100–200 feasible)

RO/R/ RE

Yes

1, 3

25/50 (100–200 feasible) 17–47 (Three Layers Announced) 17/30 (Three Layer at 47 GB) 5

RO/R/ RW

Yes

2

RO/R/ RE

Yes

1, 3

RO/R/ RW

Yes

2

RO

Yes

1

0.13

RO/RW

Yes

1, 2

MO/86 mm

2.3

RW

Yes

2, 3

Only Optical Media Yes

2, 3, 4

DVD+/-RW DVD-RAM DVD Multi (DVD+/-RW DVD Super Multi (DVD+/-RW + DVD-RAM) Prof. Disc for DATA-1 (Sony) BD HD Player/ Recorder BD Computer Drive

5.25-inch HH

HD DVD Player/ Recorder

5.25-inch HH

HD DVD Computer Drive EVD

5.25-inch HH

MiniDisc 3.5-inch ISO MO 5.25-inch ISO MO UDO (Ultra Density Optical)

Replicated/ Phase Change 120 mm Phase Change/ 120 mm

4.7/9.4 4.7–17 4.7–17

5.25-inch HH 2.5-inch HH 3.5-inch HH 5.25-inch HH

Replicated/ 120 mm MO/64.8 mm

MO/130 mm

9.1

WO/RW

5.25-inch HH

Phase Change/ 130 mm

30

WO/RO

3, 4

www.newnespress.com

56

Chapter 3

Figure 3.1: Operation of Playback with an Optical Disc.1 technical parameters and market segments served. Optical storage is the preeminent content distribution method in use today, although it is facing increasing competition from internet downloads.

3.2 Basic Operation of an Optical Disc Drive Figure 3.1 illustrates the schematic operation of playback from optical recording media. This figure shows the essential elements of an optical drive. Optical discs have land areas and grooves imprinted on the surface that provide a signal to the playback head when the head goes off track. This signal is used by the drive servo system to bring the head back on track. Another servo system maintains focus of the laser beam in the groove (or on the land). Pits imprinted into the plastic medium during the stamping phase of the replication process cause variations in the laser light reflected off of the disc surface. As the disc 1

Modified from http://content.answers.com/main/content/img/McGrawHill/Encyclopedia/images/ CE473300FG0010.gif.

www.newnespress.com

Fundamentals of Optical Storage

57

Figure 3.2: Circuit Board of an Optical Playback Drive.2 rotates under the optical head, signals are decoded into the data that was stored on the disc. Variations in light and dark areas on the optical disc surface are detected and used to position the head on the disc (focus and tracking), as well as to read the data off of the surface. Figure 3.2 shows a circuit board from an optical drive. The typical circuit board configuration is a multi-layer board with tightly packed Surface Mount Technology (SMT) devices and large Application Specific Integrated Circuits (ASICs) to provide the electronic functions. These products use relatively inexpensive components that are manufactured in high volume. Figure 3.3 shows some of the main differences between CD and DVD technology as well as higher definition blue laser optical disc formats (Blu-ray and HD DVD). The 2

Image courtesy of Peek Inside, 2007.

www.newnespress.com

58

Chapter 3 CD λ=780 nm NA=0.50 650 MB

DVD λ=658 nm NA=0.65 4.7 GB

HD DVD λ=405 nm NA=0.65 15 GB

Blu-ray λ=405 nm NA=0.85 23 GB

Figure 3.3: Difference in Spot Size, Pit Size, and Numerical Aperture for Various Optical Drive Formats.3 laser beam minimum spot size regulates storage capacity similar to transistor densities in Moore’s law for semiconductors. Conventional CD and DVD optical drives use infrared (780 nm) and red lasers (650 nm), respectively. Infrared and red laser light has a longer wavelength than blue laser light (405 nm) used for Blu-ray and HD DVD optical drives. A shorter wavelength yields a smaller laser spot represented by a smaller Numerical Aperture (NA), which in turn permits a smaller size recorded bit and tighter track pitches, and thus higher density recording. CDs and DVDs have nominal storage capacities of 650 MB and 4.7 GB respectively. Single layer BD discs have a nominal storage capacity of 25 GB. Two layer blue laser discs provide a 50 GB capacity. Blue laser DVD technology has been demonstrated for 4-layer and 8-layer discs, yielding 100 GB and 200 GB capacities, respectively, although initial products are single-layer 25 GB per side discs.

3.3 How Data Is Organized on an Optical Disc Conventional optical discs store information by making contrast variation on a disc surface that can be detected as reflected light from a read laser as the disc is rotated beneath it. For read-only discs, such as those used for content distribution, the optical contrast is made by imprinting pits in the disc surface that create light and dark regions at the pit edges. These edges can be detected by reflecting laser light off the surface of 3

Modified from NV Philips.

www.newnespress.com

Fundamentals of Optical Storage

59

the optical disc. The pit edge transitions that represent the digital data stored on the read only optical disc are organized in a spiral track. The tracks are broken up into sectors (similar to those on hard disk drives) of data with servo, error correction, and other non-user data between the data sectors. The size of the sectors is a fixed number of user data bytes such as 2,048 bytes. The physical location of data is referred to as the physical sector. The determination of the physical location of data is the job of the electronics in the data storage device. The data location as seen by the host that the storage device is attached to is referred to as the logical sector. Logical sectors virtualize the physical sectors and there is no necessary correlation between the physical sectors and the logical sectors. This virtualization of data is very common in storage devices. Data files that are made up of consecutive logical sectors may be broken up in physical sectors that can be located anywhere on the medium. For rewritable media such as hard drive magnetic media and some optical discs, the physical and logical sectors may be quite different due to reallocated sectors (required to replace bad sectors as defects develop or due to rearrangement of sectors of data as the storage medium gets close to full). For read-only media the physical sectors and the logical sectors will be pretty close to the same, since it is usually more efficient to read consecutive sectors. Optical discs vary in storage capacity because of differences in the track and linear densities, similar to the way areal density is increased on hard disk drives. Optical discs differ from hard drives in that they can record information on layers in the depth of the plastic optical disc. These are accessible by focusing the read laser beam to different focal depths in the media where information is recorded. Thus, for instance, a Blu-ray disc can have a 25 GB capacity or a 50 GB capacity depending upon whether a single layer or double layer is used. Optical read only disc layers are made by stacking layers of imprinted plastic media together to make a thicker plastic optical disc. CDs and DVDs use red lasers, which have longer wavelengths than blue laser products such as HD DVD or a Blu-ray disc. The longer wavelength light only allows detectable contrast differences from pit-land transitions on the playback medium that are larger than is the case for shorter wavelength light. DVDs have a somewhat shorter wavelength light than CDs, so they can use smaller pits and also denser tracks. Likewise blue laser products have even smaller pits and denser tracks giving them a single layer capacity of about 25 GB. With two layers this storage capacity increases by a little less than twice due to issues associated with signal loss from optical absorption of the laser light by the thickness of the medium between the laser and the detected

www.newnespress.com

60

Chapter 3

layer. The effective limit of the layers that can be used are a function of the difficulty in manufacturing multiple layers, spherical aberration compensation, and the optical losses due to absorption when reading the different layers.

3.4 Optical Disc Form Factors The most popular optical disc drive form factor is the 5.25-inch size common to CDs and DVDs. There have been many other form factors that have been introduced, and some of these are still in use, including the mini-DVDs used in some camcorders and mini-CDs that have been popular in Japan. The vast majority of optical discs are readonly discs used for software or entertainment content distribution. Optical discs are often the cheapest and fastest way to receive fixed content, such as music or video, and some companies have explored using small form factor optical discs in portable player devices. Indeed, MD (MiniDisc) technology and even CD players are used in mobile players today. In addition to MD and the Sanyo small form factor optical storage products, DataPlay (Boulder, CO) developed a micro-optical engine drive mechanism and 32 mm diameter replicated and WO phase change discs (originally provided by Ritek and Imation). DataPlay’s Digital Media product line was designed to be a robust consumer electronics product for portable audio, still image storage, and related digital content applications. Capacity was 250 MB and 373 MB (second generation) per side and 7.8 Mbps data rate; a 650 nm red laser was used. OEM drive cost was about $100 and retail media cost for a 500 MB cartridge was $10. Data Play went bankrupt, but was purchased by an investor group and is reported to be working on near-field products. Philips announced development of a small form factor optical storage device in June 2002. Using a blue laser and 30 mm diameter phase change disc, the potential for 1 GB capacity (RO and WO modes) was demonstrated. However, the company dropped the project in November 2003 with the explanation that the business units chose not to support it. LG Electronics (Seoul, Korea) is also known to be developing RO/WO SFFD optical storage products, but no announcement has been made. Finally, Sony announced the UMD (Universal Media Disc) in late June 2004. Designed for Sony’s recently announced PlayStation Portable (PSP), the new optical storage medium uses a 60 mm diameter disc and has an RO capacity of 1.8 GB. A 660 nm red laser diode is used for reading. The shell of the disc cartridge looks similar to that used for Sony’s PDD media, but is much smaller and has a relatively large UMD logo. First

www.newnespress.com

Fundamentals of Optical Storage

61

shipments were in March 2005. Although Sony has targeted this product for gaming applications, a blue laser version of UMD could easily support a 120 mm DVD-Video capacity of 4.7 GB; this would provide an elegant new consumer electronics video product. However, UMD is losing favor with game software and movie providers. Small form factor optical discs for portable players face tough competition from flash memory and small form factor hard disk drives. Also, the increasing frequency of downloading content rather than buying it in stores on optical media threatens the growth of at least the low storage capacity optical disc market (although not Blu-ray disc or HD DVD). Optical drives also use a fair amount of power to spin the disc, and thus their use in portable devices faces some challenges. When introduced, optical storage technologies are rather expensive, and even in the consumer market, drives initially sell for $1,000 or more. This price drops rapidly with production volume and yields. Because of the history of CDs and DVDs we know that once the technology ramps up the price of drives drops by 50 percent per year for the first few years. DVD players have been sold for about $40 retail with a small profit margin. In volume, optical media drops dramatically in price. Production costs of a DVD medium are a few cents per disc. Total retail DVD disc production costs, including replication and packaging, are in the range of $1.50–$2.00 each. If such a product sells for $15, then the profit margin is pretty good for pre-packaged content. IDC estimates that the OEM price of a blue laser drive could be in the range of $50 by 2010, so retail prices of less than $100 are possible, putting these products in the mainstream market.

3.5 Optical Product Reliability Optical storage products are complex. Fortunately, each new generation is able to build on its successor. Moreover, CD, DVD, and blue laser disc products are now and also in the future will be assembled from only a few subassemblies. The Optical Pickup Unit (OPU) and all mechanical elements (including the spindle and load/unload motors) are preassembled into what is called an Optical Mechanical Assembly (OMA). An electronics main board handles the servo, read/write channel, and controller functions. Final assembly consists of attaching the OMA to the chassis, attaching the main board to the OMA, attaching connectors, and statistical burn-in and test. DVD Video players, for example, use only a DVD OMA, and assign all control, electronic, and electrical functions to a system controller (SYSCON), which is

www.newnespress.com

62

Chapter 3

essentially a small, application-specific computer system. Philips and a number of Asian companies (e.g., Sanyo, Sharp, and Sony) sell kits that permit design and assembly companies to both customize and rapidly bring to market CD and DVD products (the same is eventually expected for blue laser disc products). This approach has greatly improved quality and reliability over the past 10 years. This is because complex assembly functions, for example, optical head alignment, are completed by professionals in well equipped assembly areas. Optical data storage products are amazing for several reasons: (1) a high degree of complexity reduced to ordinary assembly procedures; (2) extremely low prices relative to performance and function; and (3) a surprisingly high level of reliability. Some studies of customer service/repair records indicate that less than 0.1 percent of optical storage products have a hard failure, if they survive infant mortality. With reasonable treatment and usage, most optical storage products will give good service for at least five years. MTBF for all optical storage products is estimated to be over 100,000 power-on hours. Optical storage reliability should not be confused with return rate, which is in the range of three to five percent. This is mainly a buyer’s remorse issue, supported by easy return policies of many retailers (especially in the U.S.). Also, much emphasis has been given to optical media quality and reliability. Several reports suggest that the archival life of recordable discs is only three to five years, and magnetic tape should be used for archiving. This may in fact be true for the optical media bought for $0.05 per disc; cheap discs are only one level removed from rejects.4 High quality media produced by companies such as Maxell, MAM-A, Taiyo Yuden, and Verbatim can be expected to last at least 25 years, if not abused.

3.6 Holographic Recording Lasers with wavelengths shorter than that of blue lasers (405 nm) are possible, but it is becoming more and more difficult using known technology. Thus, other methods to increase optical recording storage density are needed besides decreasing the wavelength of the light used to record and read. Holographic recording uses a very different 4

Although not often discussed in public forums, major optical media manufacturers such as Ritek and CMC Magnetics have the ability to produce discs that fall into A, B, and C categories. A Grade is highest quality; B Grade is intermediate quality; and C Grade is lowest quality (often used as market loss leaders under obscure brand names).

www.newnespress.com

Fundamentals of Optical Storage Storage Medium

63

Data to Be Stored

eam

nce B Refere

Spatial Light Modulator

Data Pages Laser

Figure 3.4: Writing Information on a Holographic Storage Medium.5 approach to record and play back data. Figure 3.4 shows how information is written on a photosensitive recording medium by mixing a “reference” laser beam and a “data” beam. The data beam is created by means of modulation with an electro-optical device called a spatial light modulator (SLM), which forms a “page” of digital data. The mixing of the reference and data beams form interference fringes throughout the volume of the storage medium that is an analog encoding process. The recorded data are reconstructed by illuminating the storage medium with only the reference beam (now called the “reconstruction” beam). Figure 3.5 illustrates hologram reconstruction of a page of data. Two major advantages of holographic storage are massively parallel write and read and three-dimensional access to the storage medium. Rather than being recorded or read back as a stream of bits, holographically recorded data are read back in “pages” of data that can be several MBs in size (typically, 1 K × 1 K today; 2 K × 2 K and 4 K × 4 K are feasible). This, in principle, permits write and read data transfer rates to be very high. For serial streaming applications such as digital video playback, a fast FIFO buffer with at least two pages of data capacity is required to supply data to be written to 5

Image courtesy of InPhase Technology.

www.newnespress.com

64

Chapter 3

Detector am nce Be Refere

Laser

Recovered Data

Figure 3.5: Reading Information from a Holographic Storage Medium.6 or read back from a holographic disc. The ability to exploit the third dimension of the storage medium is often referred to as (hologram) “stacking.” Under ideal conditions, more than 2,000 holograms can be stacked in a fixed volume of the hologram storage material. Hence, storage densities exceeding 1 TB/in2 can theoretically be achieved. Until 2006 holographic recording was only a laboratory demonstration product. However, InPhase Technology (Longmont, CO) introduced a 300 GB Write-Once 5.25inch holographic disc with read-back speeds of 20 MBps intended for archival applications. The company believes that the technology can be extended to produce rewrite/read media and to increase storage capacity on the same format up to at least 1.6 TB with 300 MBps data rates. InPhase also believes that an inexpensive holographic distribution medium can be created. Technologies are being developed that could make holographic media an inexpensive content distribution media as well, like CDs and DVDs are today. One such concept is the holographic card (called an HROM by InPhase). An initial idea for making a linear, mass-produced playback device is the Holographic Read Only Memory (HROM) shown in Figure 3.6. The company projects that 24 GB can be stored in a 25 × 25 mm area, and on an area the size of a credit card 120 GB can be stored. 6

Image courtesy of InPhase Technology.

www.newnespress.com

Fundamentals of Optical Storage

65

Optical Module

Media

X-Y mechanism

Figure 3.6: H-ROM Concept from InPhase Technology.7

3.7 Optical Disc Storage Development Capacity and throughput for optical storage will continue to improve, though it occurs more slowly than for magnetic storage. Optical disc capacity increases not incrementally, as is the case for magnetic disc and tape, but in distinct leaps often separated by years (for example, 650 MB audio CD to 4.7 GB video DVD to 25 GB high definition blue laser discs). Moreover, optical media are removable, which in itself mandates more conservative capacity targets. Throughput for optical storage devices, as defined by data rate and access time, is well behind that of the magnetic disk. However, for the mainstream consumer electronics applications of optical disc hardware and media (CD, DVD, and most recently Blu-ray disc and HD DVD), this has little consequence. The media and supporting drives are designed for specific applications (music and video playback being the best known). Only when these optical storage products are adapted for computer applications does throughput become an issue. Optics, laser diodes, servo controls, media manufacturing quality, coding, and read/write channels have all improved greatly over the last five years. An important example is the blue (also blue-violet) laser diode. In the early 1990s, operation of GaN-based blue laser diodes outside the laboratory was not thought feasible. However, by 1995 the Japanese 7

Image courtesy of InPhase Technology.

www.newnespress.com

66

Chapter 3

chemical company Nichia (Tokushima, Japan) demonstrated the first stable devices. By 2001, the company was selling 405 nm, 5 mW laser diode kits for $5,000. Today, 405 nm 35 mW laser diodes sell for less than $50 OEM. Laser diodes with continuous wave (CW) output power of 300 mW have been demonstrated (Samsung). Moreover, operating life now exceeds 10,000 hours in many cases. Blu-ray discs (BD) and HD DVD are successor technologies to DVD for video playback and recording and PC/workstation data storage applications. They are application-specific products developed for high definition television (HDTV) playback and recording, but also can be used for computer data storage devices, as was the case for CD and DVD. HDTV requires about four times the bandwidth of standard definition television (SDTV), implying the need for four times the capacity (this actually depends on the compression algorithm used). For BD, some fine tuning of the MPEG-2 codec, 25 GB capacity (single layer), and 36 Mbps data rate yields 135 minutes of HDTV, plus extras (this contrasts with 4.7 GB and 11 Mbps for the DVD SDTV standard). Both BD and HD DVD support WriteOnce, Read-Only and Rewritable discs. They can also use multilayer discs to increase the storage under the optical head. Capacity is limited by optical absorption, spherical aberration, and SNR of the optical signals through the layers. To achieve HDTV level capacities and data rates, a blue laser diode must be used for this technology. Additionally, an objective lens with a higher NA is required. The BD approach is (1) to implement “near front surface” recording (a 0.1 mm protective layer and hard coat combination is used to protect the data layer or layers), which increases tilt and aberration tolerances and (2) to use a NA = 0.85 objective lens. HD DVD preserves the basic 2 (polycarbonate) substrate design of DVD and uses a NA = 0.65 objective lens; active tilt control is probably also used. At the time of this writing Blu-ray disc and HD DVD are engaged in a market war to see which product can achieve de facto standard status for the consumer industry. Blu-ray disc is promoted by Sony, Philips, Panasonic, and numerous partners (about 178). HD DVD is promoted by Toshiba, NEC, and other members of the HD DVD Alliance. At the 2007 CES, LG announced an optical drive that can read both formats. Such products, if they become common, could help considerably in making consumers comfortable with choosing one format or the other for their HDTV viewing, knowing that they will still be able to read whatever format does not win the format war.

www.newnespress.com

Fundamentals of Optical Storage

67

Table 3.2: Comparison of DVD, Blu-ray Disc, and HD-DVD DVD

Blu-ray Disc

HD-DVD

Capacity Data Transfer Rate (Mbps) Laser Cover Layer Multi-layer potential Video Compression Film Industry Support

4.7–8.5 GB 11

25/50 GB 36

15/20 GB 36

Red 0.6 mm 2 max

Blue-violet 0.1 mm 8 max

Blue-violet 0.6 mm 3–4 max

MPEG2

MPEG2, WM9

Supporting Companies

DVD Forum (Over 200 Companies) DVD Forum Steering Committee (20 Companies)

MPEG2, MPEG4, WM9 Sony Pictures, Walt Disney, 20th Century Fox, MGM, Paramount, Warner Bros. Blu-ray Disc Association (Over 160 Companies) Apple, Dell, HP, Hitachi, LG, Panasonic, Mitsubishi, Pioneer, Philips, Samsung, Sharp, Sony, TDK, Thompson, Walt Disney, Warner Bros., 20th Century Fox

Core Members

All

Universal, Paramount, New Line Cinema, Warner Bros. HD DVD Promotion Group (60 Companies) Toshiba, NEC, Sanyo, Memory-Tech

Table 3.2 compares some of the important characteristics and support base for Blu-ray disc and HD DVD compared to DVD technology. Figure 3.7 gives projections for the shipment of various optical drive formats used for consumer electronics applications through 2011. Until the format war is resolved and the price of hardware and media drops to the point that these blue laser products become mainstream products (2011 in this projection), conventional red laser DVD and CD drives will dominate. This projection shows negative overall drive growth after 2007, probably due to the expected growth in popularity of downloading versus content distribution with optical media. Whether this happens depends upon continued growth in the availability of high bandwidth internet access.

www.newnespress.com

68

Chapter 3 180.0

Blue Laser Players/Recorders

Millions of Drive Units

160.0

Red Laser Players/Recorders

140.0 120.0 100.0 80.0 60.0 40.0 20.0 0.0 2006

2007

2008

2009

2010

2011

Years

Figure 3.7: Projections for Consumer Electronics Optical Drive Shipments.8

Increasing storage capacity of optical discs tends to come in discrete and significant jumps to support higher definition content formats. For instance 650–750 MB CDs are used for single music album distribution while 4.7 GB DVDs are used for SD movie distribution. 20–50 GB blue laser discs enable HD movie distribution. If, as expected, physical content distribution continues to be the main use of optical discs in the future, then the growth of optical disc capacities and the required technologies will follow the growth in home (and professional) content resolution. Figure 3.8 shows a projection for the required data rate and storage capacity to contain various sizes of multimedia objects. As can be seen from the figure, blue laser optical discs will be large enough to contain compressed HDTV quality movies, but higher resolution compressed content will require hundreds of GBs, maybe beyond the 200 GB maximum discussed with multilayer blue laser products. It is likely that in the early part of the next decade resolutions as high as those currently used in digital cinema (4 K × 2 K pixel; HDTV is 2 K × 1 K pixel) could begin to appear in some high end households and could be commonplace by about 2020. In addition humans want ever more realistic video experiences that require even higher resolution and depth requirements. Imagine for instance the storage requirements for a 3D movie displayed in full depth at 4 K resolution in some sort of home projection system viewable from every direction. Such products could well become 8

Courtesy of iSuppli, 2007.

www.newnespress.com

Fundamentals of Optical Storage

69

rg et A

Professional HDTV-Quality Movie (Compressed)

Ta

1,000

re a

Digital Cinema and VR

100 Next Gen DVD Technology

10

One Page ASCII Text

0.01 1 KB

10 KB

DSP Scanned b/w Voice Page (300 pix/in)

100 KB

1 MB

VCR-Quality Movie (MPEG-1)

10 MB 100 MB

1 GB

DVD Technology

0.1

CD Technology

Next Gen DVD Technology

1

DVD Technology

Broadcast-Quality Movie (MPEG-2) CD-Quality Stereo Audio

CD Technology

Data Rate (Mbps)

Upper Limit for Conventional Optical Technology

10 GB

Consumer HDTV-Quality Movie (Compressed)

100 GB

1 TB

Multimedia Object Size

Figure 3.8: Comparison of Storage Capacity Requirements and Data Rates for Different Size Multimedia Objects.9

common household products within the first thirty years of this century. The result would be demand for optical physical distribution media with storage capacities of 1 TB or higher. Discs of 1 TB or larger could be a possible role for mass produced holographic discs, large numbers of optical layers on a disc, or for even higher frequency optical disc products. Even with the rise of content distribution through the internet, as long as the resolution requirements increase faster than the bandwidth available to most consumers, there will continue to be demand for physical media. The mass production cost of optical media tends to be a few cents per disc. As the storage capacity increases, the cost per unit of storage becomes very low. As we move from DVD to blue laser optical discs, the price per GB consequently decreases. The $/GB of optical media will continue to drop if holographic recording or even higher frequency optical recording moves into consumer products. 9

Used with permission of Telcordia.

www.newnespress.com

70

Chapter 3

3.8 Chapter Summary •

Optical disc storage technology has been around for over 35 years. There are many optical storage products now available on the market. Some of these are professional products for niche markets, while others are designed for consumer applications. Many of the consumer optical formats are primarily used for content distribution of music or video.

•

Optical disc drives use mirrors and lenses to move laser light around and focus it on the target optical disc for reading or writing. As the wavelength of the laser light decreases, the optical spots that are written or read become smaller and so higher density recording is possible. Thus blue laser optical discs can record more data than red laser optical discs.

•

Data is written on optical discs as pits on predefined tracks. The optical contrast of the transition from the pit edges provides the signal that contains the digital information. Much like a hard disk drive the data on the tracks is broken up into sectors with data and servo information in each sector.

•

Although the 5.25-inch optical disc format is the most popular for CD, DVD, and blue laser DVD products, other form factors have been used in the 35-plusyear history of optical recording. Smaller form factor discs have been popular in players and in video recorders, especially in Japan. DataPlay and some other players have tried to introduce truly portable form factor optical discs but so far such products have not developed into a mature market.

•

Optical discs can have a range of expected lifetimes depending upon the care and materials with which they have been made. High quality optical discs kept in an ideal environment (temperature and humidity) can be expected to survive for 20 years (probably long after the devices that could read them have disappeared). Optical storage disc drives are generally specified for 100,000 power-on hours.

•

Holographic recording can record information in an optical disc format through the volume of the recording media. This technology, although worked on for over 30 years, has never come to production. Today companies like InPhase are trying to bring optical recording to market.

•

Rich content distribution drives the development and use of new optical recording technologies. Blue laser optical discs of the HD DVD and Blu-ray

www.newnespress.com

Fundamentals of Optical Storage

71

schools are currently fighting for overall market dominance. Many users are waiting to see which technology will win before they make their purchase commitments. Beyond the current battle for blue laser formats to support HD content distribution, new resolution formats are in process that could require even larger disc formats than 25 or 50 GB. Optical disc products, perhaps using holographic or some other advanced optical storage technology, could someday provide 1 TB optical discs for very high resolution video content distribution.

www.newnespress.com

This page intentionally left blank

CHAPTE R 4

Fundamentals of Flash Memory and Other Solid State Memory Technologies

Objectives in This Chapter •

Review the history of flash memory technology.

•

Describe the basic flash operations of erase, write, and read.

•

Explore the causes and operation of flash memory cell wear.

•

Present the differences between NOR and NAND flash memory technologies.

•

Understand bit errors in flash memory.

•

Find out how to reduce the negative effects of flash cell wear with wear leveling.

•

Discover the basics of bad block management on flash memory.

•

Review embedded versus removable flash memory technology.

•

Learn how flash file systems make flash data accessible.

•

Find out about the advantages and trade-offs of single level cell memory and multilevel cell memory.

•

See how die stacking can be used to increase volumetric density of flash-based devices.

•

Explore the various ways that flash reliability can be ensured and what sort of useful life can be expected in an application.

www.newnespress.com

74

Chapter 4 •

Learn the various flash memory card formats.

•

Understand what can be expected in terms of flash memory prices over the next few years.

4.1 Development and History of Flash Memory Flash is an extension of the floating gate method of manufacturing nonvolatile memory. The first sort of floating gate memory was the Erasable Programmable Read Only Memory (EPROM), invented in the 1960s but not developed until the 1970s. In EPROM, like in Dynamic Random Access Memory (DRAM) and Read Only Memory (ROM), each memory bit was represented by a transistor. Appendix C gives a list of some of the most significant flash memory manufacturers. It helps to understand how transistors work if you want to understand these circuits. In a Field-Effect Transistor (FET), the FET current flows from the source to the drain. The gate controls how much current flows through the channel (the area between the source and drain). If the gate is unbiased, then the current flows through the channel relatively freely. If a bias is applied to the gate, then the channel depletes, that is, the carriers are moved out of part of the channel, making it seem narrower, and limiting the current flow. This principle is key to all FET-based technologies, Metal-Oxide Semiconductor (MOS), Complementary Metal-Oxide Semiconductor (CMOS), Bipolar Complementary Metal-Oxide Semiconductor (BiCMOS), and so on. The ability to turn a circuit’s current flow on and off allows individual bits to be routed to a data bus. In a DRAM each memory bit’s transistor uses a capacitor to store the bit. The case is different with a ROM, where each bit’s transistor uses a short or open circuit to represent a bit (either programmed at manufacture by a mask, or afterward by a fuse that could be blown). In EPROM, the transistor itself looks as if it contains something kind of like a capacitor, but it is actually a second gate complementing the control gate of each bit’s transistor. The bit’s transistor actually has two gates, one that is connected to the bit line, and one that was connected to nothing—it “floats.” The technology is known as “floating gate.” This concept is used by four technologies: EPROM, Electrically Erasable Programmable Read Only Memory (EEPROM), NOR, and NAND. Figure 4.1 illustrates the cross section of a floating gate. When a bit is programmed, electrons are stored upon the floating gate. This has the effect of offsetting the charge on the control gate of the transistor. If there is no charge upon the floating gate, then the

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

Source

Control Gate Gate Oxide Floating Gate Tunnel Oxide

75

Drain

Channel Substrate

Figure 4.1: Floating Gate Flash Memory Cell. control gate’s charge determines whether or not a current flows through the channel. A strong charge on the control gate ensures that no current flows. A weak charge will allow a strong current to flow through. The floating gate is capable of storing a charge of its own. This adds to the bias of the control gate. If there is a charge on the floating gate, then no current will flow through the channel, whether or not there is a charge on the control gate.

4.2 Erasing, Writing, and Reading Flash Memory The key to using floating gate technology is to put a charge onto the floating gate when it is needed, and to take it off when it is no longer needed. There are two methods of getting charges onto and off of the floating gate: Channel Hot Electron injection (CHE) and Fowler-Nordheim tunneling (FN). Both concepts involve quantum methods which are beyond the scope of this text, so they will be dealt with here in a very simplistic way. Channel hot electron injection takes advantage of the fact that electrons are more mobile with high current levels. When CHE is used to add electrons to a floating gate, a high source to drain current is run through the channel. With this high current, a number of electrons are looking for ways to “boil” out of the channel. A positive bias on the control gate attracts electrons from the channel into the floating gate, where they become trapped. This adds a bias onto the floating gate (see Figure 4.2). Fowler-Nordheim tunneling requires a high voltage (compared to the operating voltage of the chip—usually between 7–12 V) to be placed between the source and

www.newnespress.com

76

Chapter 4

High Current

Attracting Charge

+

Some Electrons “Boil Off” and Drift into Gate

- - - - - - - -

Figure 4.2: Channel Hot Electron Injection.

Strong Negative Potential

Strong Positive Potential

---

+++

- - - -

Electrons “Tunnel” Through Gate Oxide

---

Strong Negative Potential

Figure 4.3: Fowler-Nordheim Tunneling.

the control gate of the transistor. If the voltage is sufficient, the electrons “tunnel” through the gate oxide layer and come to rest upon the floating gate. This is illustrated in Figure 4.3. In EPROM, the original floating gate technology, the memory was erased through exposure to ultraviolet light. Light rays would allow the electrons trapped in the floating gate to migrate back to the channel. In EEPROM, Fowler-Nordheim tunneling was reversed by a second transistor on the memory bit to take electrons back out of the floating gate and put them back into the source where they originally came from. This erasure approach is illustrated in Figure 4.4. Both NAND and NOR flash technology use the same electronic erasure concept as EEPROM, but rather than add an extra erase transition to each individual bit’s transistor, a single very large transistor erases all the

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

Strong Positive Potential

Strong Negative Potential

+++

---

- - - -

Electrons “Tunnel” Back Through Gate Oxide

+++

77

Strong Positive Potential

Figure 4.4: Erasing the Bit Through Tunneling. transistors in a sub-array called a sector. This effectively halves the size of the memory array, providing significant cost savings to designers who do not need to be able to individually erase each bit or byte of the memory.

4.3 Difficulties that Cause Wear in Flash Memory Tunneling electrons migrating through the tunnel oxide sometimes cause difficulties. It is inevitable that electrons from time to time will get trapped in the tunnel oxide. These electrons, once trapped, cannot be removed. This will impact the operation of the cell to a certain degree, depending upon the number of electrons trapped in the oxide: if a lot of electrons are trapped then there will be a big impact, but a low number of electrons are not likely to cause much impact at all. As the number of trapped electrons in the oxide gets too large, it becomes harder to maintain electrons in the floating gate. Thus the memory cell leaks the stored charge. The number of electrons trapped in any one cell’s tunnel oxide is a function of how the chip is made, and more importantly, of how many times that particular transistor has been erased and rewritten. A specification has thus been devised to recommend a maximum number of erase/write cycles that a memory cell on a certain chip can withstand before a failure is likely to occur, and this specification is called the chip’s endurance. Typical endurances have been managed up to very high levels over the course of time. For many chips today, endurance is specified as 105 erase/write cycles. As we will later see, some novel approaches are used to further remove the possibility that endurance failures will impact the operation of a system.

www.newnespress.com

78

Chapter 4

4.4 Common Flash Memory Storage Technologies: NOR and NAND There are two kinds of flash memory, NOR and NAND. The two terms are names of types of logic gates, the negated OR function and the negated AND function. The big difference between the two types of architectures is real estate. NAND has a significantly smaller die size than does NOR. This translates to significant cost savings. These cost savings come with a trade-off. NAND does not behave like other memories. While NOR, SRAM, and DRAM are random access devices (the “RAM” part of DRAM and SRAM stands for random access memory) NAND is part random and part serial. Once an address is given to the device, there is a long pause, then that address and several adjacent addresses’ data come out in a burst like a machine gun. We will explain why in the next two sections. Another way that NAND makers squeeze out costs is by taking some shortcuts in ensuring data integrity. NAND was designed to replace hard disk drives back in the 1980s. At that time some thought that there was a brick wall limiting the capacity of magnetic storage. Once hard disk drives hit that brick wall then semiconductor memory could race past it in density and cost. NAND’s inventors borrowed a leaf from the hard disk drive designers’ book by allowing their design to have occasional errors that would be corrected by external logic. Both hard disk drive designers and NAND designers were able to use this more relaxed approach to data integrity in order to squeeze more bits onto a given piece of real estate (whether on a magnetic disc or on a silicon chip) than would be possible if absolute integrity were to be maintained. Two techniques, serial access and lower data integrity, allow NAND die sizes to be less than half the size of their NOR counterparts. Let’s look at the two of these in depth.

4.4.1 How Does NOR Memory Work? We will use a classroom analogy. Picture a class where there is space around all the desks. Each of the desks will represent a bit transistor in the memory, with the students representing data. The teacher represents the system interface that handles the data transaction between the rest of the system and the memory bits.

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

79

T

A1

B1

C1

D1

E1

F1

G1

H1

I1

J1

K1

L1

M1

N1

A2

B2

C2

D2

E2

F2

G2

H2

I2

J2

K2

L2

M2

N2

A3

B3

C3

D3

E3

F3

G3

H3

I3

J3

K3

L3

M3

N3

A4

B4

C4

D4

E4

F4

G4

H4

I4

J4

K4

L4

M4

N4

A5

B5

C5

D5

E5

F5

G5

H5

!5

J5

K5

L5

M5

N5

Figure 4.5: Classroom Like NOR—A Spacious Layout.

We have labeled the students in Figure 4.5 as A1, B1, and so on, to indicate the row and column where that student sits. In this example there are 14 columns (A–N) and five rows (1–5), giving 70 students all together. Most memories are designed similarly to this classroom. There are rows and columns, and word lines to tell the memory bits which row is being requested and bit lines to access a column within that row. When a memory bit is requested, it is similar to the teacher calling a student to the front of the room. Say the teacher in this example called for student B3 to come to the front of the room (Figure 4.6). That student would simply walk from his or her desk up the aisle to go to the teacher’s desk. All students could get to the teacher’s desk in about the same amount of time, and it would not take very long to get there.

www.newnespress.com

80

Chapter 4

T B3

A1

B1

C1

D1

E1

F1

G1

H1

I1

J1

K1

L1

M1

N1

A2

B2

C2

D2

E2

F2

G2

H2

I2

J2

K2

L2

M2

N2

C3

D3

E3

F3

G3

H3

I3

J3

K3

L3

M3

N3

A3

A4

B4

C4

D4

E4

F4

G4

H4

I4

J4

K4

L4

M4

N4

A5

B5

C5

D5

E5

F5

G5

H5

!5

J5

K5

L5

M5

N5

Figure 4.6: The Teacher Calls Student B3 to the Front of the NOR Classroom.

This is very similar to the way that a NOR memory works. All data can be accessed rapidly, and all accesses take about the same amount of time. Data can be requested from individual locations in a completely random sequence.

4.4.2 How Does NAND Memory Work? Now let’s say that some cost-cutting measures were implemented by this school. The administration found that they could sell one of their buildings if they used their existing buildings more efficiently. Their plan is to shrink the size of each classroom, which they can do if they push all the desks up against each other, with the desks on one side of the classroom pushed right against the wall. This is illustrated in Figure 4.7. It is clear that there is significant space savings by taking this approach.

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

81

T

A1 B1 C1 D1 E1 F1 G1 H1 I1

J1 K1 L1 M1 N1

A2 B2 C2 D2 E2 F2 G2 H2 I2

J2 K2 L2 M2 N2

A3 B3 C3 D3 E3 F3 G3 H3 I3

J3 K3 L3 M3 N3

A4 B4 C4 D4 E4 F4 G4 H4 I4

J4 K4 L4 M4 N4

A5 B5 C5 D5 E5 F5 G5 H5 !5

J5 K5 L5 M5 N5

Significant Savings!

Figure 4.7: In the NAND Classroom All the Desks Are Pushed Together, Saving Valuable Space.

This approach does not come for free, though. Now, when the teacher calls student B3 to the front of the room, all the students in B3’s row must get up and walk to the front of the room in a line (Figure 4.8). There is simply not enough space for them to do it any other way. Once the row is in front, then B3, or C3, or N3, or any of these students can get to the teacher rapidly. On the other hand, none of these students can reach the front of the room until all the students blocking the way have gone to the front. This is similar to the way that NAND is laid out. Just as the students share a single aisle at the far side of the classroom, many bits in NAND share a bit line, dramatically reducing the amount of space used on the NAND die to move data back and forth to the bit transistors. Like the classroom analogy, NAND takes a longer time to get a memory bit (a student) that is randomly called to the rest of the system (the front of the class). Once that bit line’s data is ready (the student’s row is

www.newnespress.com

82

Chapter 4

T B3 N3 M3 L3 K3 J3

I3 H3 G3 F3 E3 D3 C3 B3 A3

A1 B1 C1 D1 E1 F1 G1 H1 I1

J1 K1 L1 M1 N1

A2 B2 C2 D2 E2 F2 G2 H2 I2

J2 K2 L2 M2 N2

A4 B4 C4 D4 E4 F4 G4 H4 I4

J4 K4 L4 M4 N4

A5 B5 C5 D5 E5 F5 G5 H5 !5

J5 K5 L5 M5 N5

Figure 4.8: In the NAND Classroom All the Students in a Row Must Move for One of Their Row-Mates to Get to the Teacher.

in the front) then data from that word line can be presented to the system at a rapid-fire rate (just as every student in that row can get to the teacher’s desk very quickly). From this it should be clear that NAND is less expensive to manufacture than NOR, since the manufacturing cost of a silicon chip is a function of its size. Keep in mind, though, that the cost savings might be offset by the two technologies’ significant difference in functionality. Although NOR usually seems easier to use, surprisingly enough, this difference in functionality actually makes NAND a better fit than NOR in applications where serial accesses are preferred to random accesses. Video and audio streaming are two very good examples of applications where serial access is preferable to random.

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

83

4.5 Bit Errors in NAND Flash As mentioned earlier, NAND was designed to replace hard disk drives. The inventors realized that the hard disk drive’s innovators had come up with a lot of good ideas, and borrowed some of those ideas to further drive costs down for NAND. One very important similarity is the handling of bit errors. The serial technology in NAND is very prone to bit errors, something that digital systems cannot tolerate. In order to get past this, NAND makers used error correction technology common to the hard disk drive business, for instance a 256-bit sector will have an additional 16 bits added on. These check or parity bits are programmed with a highly compressed code that indicates what the data in those other 256 bits should look like. This implies that every time that a sector is programmed, then corresponding parity bits must be calculated and stored along with the original bits. A controller is usually given this task. This is one of the reasons that NAND most often ships coupled with a controller chip, although there are cases where the functions of the controller are embedded within another processor in the system. During a read, the data bits in a sector cannot always be trusted, so they are never simply read from the NAND into the system. Instead, the 256 bits in the sector are jostled with the 16 parity bits in an error detection and correction engine, which presents a sector’s worth of corrected data to the system. Different controller manufacturers use different algorithms in their controllers, resulting in different degrees of correction. You will hear terms like “four-bit correct, six-bit detect” to describe the capabilities of a controller. The higher the number, the more complex the algorithm required. This means that a more expensive CPU is often necessary to perform that processing in a reasonable period of time. Fortunately, processing power’s cost decreases over time, so better and better algorithms are likely to be used in even the lowest-cost systems.

4.6 Managing Wear in NAND and NOR NAND and NOR memories are treated differently when it comes to managing wear. In many NOR-based systems, no management is used at all, since the NOR is simply used to store code, and data is stored in other devices. In this case, it would take a near-infinite amount of time for wear to become an issue, since the only time the chip would see an erase/write cycle is when the code in the system is being upgraded,

www.newnespress.com

84

Chapter 4

which rarely if ever happens over the life of a typical system. In other applications where NOR is used to store data (for example the large NOR chips used by numerous cell handsets), major NOR suppliers provide file management software that implements wear leveling algorithms that will be explained in detail when we discuss NAND below. NAND is usually found in very different applications than is NOR. Whereas NOR is most often used for code storage, NAND is most often used for data storage. This means that the end user will continually be erasing and rewriting the data in the NAND. Although 105 erase/write cycles is a very large number, NAND makers still treat loss of data very seriously and manage the wear in the NAND to ensure that a cell runs the lowest possible risk of wearing out. As mentioned above, NAND is always connected to the system by way of a controller. In some cases, this controller is a dedicated chip. Examples of this are most flash cards and all USB drives. When a customer buys one of these, he or she is buying one or more NAND chips plus a dedicated controller. In other systems the functions of the controller are implemented within some other processor, perhaps the main processor of the system. In cell phones with embedded flash, for example, the controller is implemented as a function of the cell phone’s baseband processor. One of the responsibilities of this controller is the implementation of the wear leveling algorithm. The controller, by being a gatekeeper, can disguise the layout of the NAND chip from the rest of the system. If the system requests the data from address “A,” the controller can present a very different address “B” to the NAND, and provide this data to the system. This will be OK if the last time the system wrote to “A,” the controller forced that data into location “B.” The controller keeps track of where everything is physically located, and where it looks like it is located to the system, in a table. This is a relatively simple task. Another feature of this table is that it tracks how often erase/write cycles have been performed upon each erasable sector of the NAND chip. This way, the controller can ensure that all sectors get an even number of erase/write cycles, thus spreading the wear across the chip. Let’s say that a user keeps changing the fifth song in their MP3 player, but is satisfied with all the rest of the songs. The controller will notice that the space initially used for the fifth song is getting more erase/write activity than other parts of the chip, and will decide to move the actual location in the flash memory where the fifth song is stored. If there is free memory, the fifth song will be moved to the free area. If there is no free memory, then something that is rarely changed is likely to get moved to

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

85

the fifth song’s old space, and the fifth song will be moved to that part of the NAND. It is likely in such a case that sections of the NAND that are never written to keep moving around, since they sit in prime areas with a very low number of erase/write cycles.

4.7 Bad Block Management Along with wear leveling, the controller in a NAND-based system is responsible for bad block management. The controller, as a part of its data correction responsibilities, keeps track of how much error correction is performed on each sector of the NAND. If the number reaches some predetermined high level, that sector is marked as “bad,” and will no longer be used by the controller. This has a very unusual impact upon the user, one that is rarely seen in any digital system: graceful degradation. When most digital systems experience a problem, they come to a sudden stop. The screen goes blank, or the On/Off button stops working. In NAND with bad block management, as blocks get taken out of the system, the overall size of the NAND slowly diminishes. In an extreme case, for example, if you had a very small flash card in a system and continually erased photos to snap new ones, you would notice that your card first stored 25 photos, then one day it only stored 24. Later it would slip to 23, then 22, all very gradually. You wouldn’t lose any pictures, but the number of pictures you could store would gradually decline.

4.8 Embedded Versus Removable NAND Flash Some systems use removable flash memory cards where others have the flash memory soldered into the system. What is the difference between how these systems work? In reality, the differences between these two kinds of systems are relatively minor. The functions of the controller in a flash card are usually implemented using firmware in an embedded processor. In systems with cards, the main processor of the system is not burdened with error correction, wear leveling, or bad block management, as this is the responsibility of the card’s controller. In embedded systems, the wear leveling, bad block management, and error correction are taken care of by the main processor. As we will see later, some cards (SmartMedia and xD-Picture cards) do not use a controller, offloading these functions to the main processor of the system the card is plugged into.

www.newnespress.com

86

Chapter 4

4.9 Flash Memory File Systems Like a hard disk drive, the flash contains files, and those files must be managed. This is in addition to the error correction, bad block management, and wear leveling discussed above. This is something that designers need to keep in mind, as many firmware designers who start with a small piece of code realize that their system can expand to the point that formal file management software could simplify their data management chores. File management software is widely available today, and this software has been modeled on that used for decades by hard disk drives. In effect, this software hides the difference between the NAND and a hard disk drive for the host device, so that software that was written for NAND could be relatively easily ported to a hard disk drive based system and vice versa.

4.10 Single Level Cell and Multilevel Cell Flash Memory As we saw earlier, the manufacturing cost of a flash chip is a function of the area of the chip. For cost’s sake, it is important to get the chip as small as possible. One unique way to do that is to store two bits of data in the area where one bit would normally go. This is something that has been employed by flash makers, but has not appeared in other forms of memory. To understand this concept, consider the basic flash memory cell in Figure 4.1 as a linear device, rather than as a digital one. The number of electrons stored upon the floating gate is proportional to the time spent charging it, giving the user some control over this charge. The current flowing through the channel will be proportional to the charge on the floating gate when the control gate is not blocking current flow through the transistor. One company, Information Storage Devices, made a business in the 1990s of storing linear values on similar transistors to record speech without digitizing it. But let’s stay digital. In order to store two bits onto a transistor, the four different states represented by those two bits must each be represented (00, 01, 10, and 11). This can be done by storing four voltage levels. Instead of storing a high charge on the floating gate for a logical 1 and a low charge for a logical 0, using the halfway point as the decision level between a 1 and a 0, four levels can be stored, nothing, one third, two thirds, and

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

87

full, each representing one of the four states. The circuitry for discerning the difference between these levels is a bit more complex, but this technology is well understood. Chips that store one bit on a cell are usually called SLC for single level cell chips. Those that store two bits on a cell are called two level MLC for multilevel cell chips. The most important issue in making MLC chips is pulling the data out of the device. When a transistor is read, the reading circuit has to discern between four small voltage levels rather than two big ones. Digital chips are noisy critters, and the more noise there is in a system, the more difficult it is to discern small voltages apart from one another. This challenge has caused leading flash suppliers a good share of headaches. One part of the solution is to give the chip time for the noise to settle down before making a decision. This slows the chip down, which is why MLC chips are usually slower on read cycles than their SLC counterparts. Somewhat similarly, writing into these chips is a bit more daunting than writing into SLC chips. Across the chip, different cells behave differently, and relatively sophisticated state machines (tiny little computers) are used first to put a smidgen of charge onto a floating gate, next to measure to see if the floating gate’s charge is near the optimum charge for the cell to represent one of the four levels, then to continue iterating this process until it is done. Usually, a smaller current is used to give better control over the charge, and this slows down programming. Precision takes time. This is a good idea. Why not take it to the next level to store more than two bits on a cell? This is precisely what several companies have tried to do for a number of years, but there are many problems that have proven tough to overcome. The biggest one is that the fractional voltages, as they get smaller and smaller, become increasingly difficult to pull out of the noise inherent on a digital chip. Another is a mechanism called “adjacent cell disturb.” This is when part of the charge on one flash cell leaks sideways into an adjoining cell. As the levels of charge on the floating gates become smaller and smaller, the impact of a few electrons migrating from one floating gate to another becomes more and more significant. A lot of work is going into solving these problems. For instance, SanDisk has been working on a three-bit NAND cell, and absorbed another company, M-systems, who planned to sample a four-bit per cell NAND solution toward the end of 2007. As if that were not enough, participants in the NAND market are already looking into ways to go beyond four bits per cell. SanDisk is once again a leader in this campaign,

www.newnespress.com

88

Chapter 4

having purchased a company called Matrix Semiconductor in 2006. Matrix developed a means of stacking transistors one atop another to get to 8 or even 16 levels, multiplying the amount of memory on a given size chip by the same number. There is a long way to go with technologies like these!

4.11 Another Approach to Multilevel Cells There is a completely different approach to storing multiple bits on a cell that is used by fewer manufacturers but is quite possibly going to find expanded use in the future. This technology goes by different names, the most common being Saifun’s NROM, Spansion’s MirrorBit, and its near-equivalent technology from Macronix. This technology actually uses a flash cell that is manufactured slightly differently than the typical cell of Figure 4.1. In the typical gate, the floating gate is made of conducting polysilicon. In the MirrorBit cell, an insulating nitride layer replaces the floating gate’s polysilicon. In such a cell, the higher tunneling voltage is applied only on the source and not on the drain. Since the floating gate is an insulator, any charge placed on the floating gate ends up near the source. This means that a different charge can be put on one side of the floating gate or the other simply by swapping the source and the drain with each other. Reading the cell is a little odd. When a current runs from the source to the drain, a zone grows around the source that is impervious to the charge on the gate. This means that the current from the source to the drain will only be impacted by gate bias that is not near the source. The current will ignore the charge on the floating gate that is close to the source, and will only be impacted by the charge near the drain. Once again, by exchanging the source with the drain, we can choose which side of the cell we want to read, the left or the right. This is a tidy means of achieving two bits per cell, and Spansion claims that the process is less expensive to manufacture than that of standard floating gate technology. The next question is where to go from here? It is clear that the same trick used to store multiple levels on a conventional floating gate can be used in a NROM cell, and this is exactly what Spansion aims to do, with devices sampling at the writing of this chapter. It would not be surprising to see such a technology further expanded with other existing four-bit schemes to make an eight-bit cell in the future.

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

89

4.11.1 Stacking Die to Achieve Higher Storage Capacity Apart from schemes to get the most storage on the least amount of silicon is a push to get the most storage in the smallest physical space within the system. This is especially important for portable devices like cell phone handsets and high-capacity flash cards and MP3 players. A common approach to this is to stack memory chips within a package that would normally be used to house a single chip. For the most part two chips are stacked one on top of the other to get double the storage in the same size package, but it is not uncommon to see stacks with three or four chips, and vendors today claim to have the capability to stack as many as nine chips in upcoming products. The stacking technology most widely used is to wire-bond chips that have been ground on the back side so they are very thin. Wire bonding is the oldest technology for connecting a die to the outside world and it is very impressive that it can be done upon multiple levels of chips as shown in Figure 4.9. Memory chips are particularly well-suited to stacking, because most systems use memories in such a way that only one chip is powered up at a time, so a hot chip will never be pressed up against another hot chip. This would not be the case with processors or other logic chips like ASICs or Application Specific Standard Products (ASSPs). Of all memory chips, NAND lends itself best to stacking because of its bit errors. This is because memory chips are very sensitive devices, and two chips that are destined for

z

Figure 4.9: Die Stacking Showing Wire Bonds at Multiple Levels.

www.newnespress.com

90

Chapter 4

stacking must first be tested to ensure that every single bit works flawlessly. Only after they have passed this test can they be stacked and packaged, whereupon they have to be tested all over again, and scrapped if they fail. Although NAND chips are just as prone to damage as are other chips, the fact that the NAND is automatically going to be paired with a controller (that performs error correction and bad block management) allows chips that might otherwise fail to still be useable. In some cases a semifunctional stack is sold with an unusual storage capacity—some non-binary number like 2.5 GB or 3 GB—because of the failures due to the stacking.

4.12 Trade-offs with Multilevel Flash Memory In the MLC versus SLC section earlier in this chapter, we mentioned the difference in endurance between SLC NAND and two level MLC NAND Flash. There is a 10-to-1 difference in endurance between the two—from 105 cycles to 104. In today’s typical NAND applications, this is unlikely to pose much of a concern. In the future, as NAND is used to replace other technologies like hard disk drives, and possibly even DRAM, this may be of significant concern. Say, for example, that a counter that is usually implemented in DRAM is instead implemented in NAND. Just to make the situation worse, let’s make this the line counter for a graphics processor. Every time the raster scans from one line to the next the counter is read from memory then incremented and rewritten. This is not at all atypical for a true DRAM application, but so far NAND has not been used in this sort of application. One very good reason is that the write speed of flash does not accommodate this kind of activity, which would require a new write every 32 µs. This is much less than the 200 µs page write time given in our prior example. Looking past that limitation, though, what would happen were we able to write to the line counter every 32 µs, and the line counter stayed at the same address in the flash? The simple answer is that 105 writes at 32 µs intervals would put the design in jeopardy of a bit failure after its first 3.2 s of use. This is nowhere near acceptable! Keep in mind that this is the SLC endurance. With MLC the system would run the possibility of failing after one tenth that time: 0.32 s! We choose this example to point out that an engineer needs to consider how he or she will use various memory storage sites to determine which is best suited for DRAM and which can take advantage of NAND.

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

91

As was explained in the prior wear leveling section, the size of the flash in a wear-levelmanaged system has a big impact upon the time it takes to enter the danger zone of an endurance-related failure. Although two level MLC endurance is 10 times less than that of SLC, the MLC capacity is usually double that of its SLC counterpart, reducing the concern by half (assuming appropriate use of wear leveling). Should a designer throw memory capacity at the problem, the difference between MLC and SLC could become academic. As for performance, MLC is slower than SLC, and this point needs to be managed carefully. One approach is to employ parallelism—when one chip is delayed by its slower operation a second (or third or fourth . . .) chip takes over. In very small systems this is not a viable approach, but if it is necessary for a system to use multiple flash chips to achieve its desired storage capacity, there is no reason that these chips cannot be operated in parallel to improve speed.

4.13 Types of Flash Memory Used in Consumer Electronics Devices Although most consumer electronics devices in the past used NAND in a card format, there is a progression away from that model as flash content increases. It appears that the flash content of earlier systems motivated users to store all of one type of data in one flash card, and all of another into another card. As one card was filled, it would be taken to the office for eventual offloading to a hard disk drive, while the other was still in use. Today many consumer electronics devices contain such an abundance of flash that this flash can be split into zones where one zone suits one purpose and another suits another purpose, with the threat of the flash becoming filled diminishing over time. This of course assumes that the required capacity of the files also doesn’t increase in time.

4.14 Flash Memory Environmental Sensitivity The key advantage that flash memory devices have over electro-mechanical storage, such as hard disk drives or optical discs, are their extreme ruggedness. From time to

www.newnespress.com

92

Chapter 4

time, stories appear attesting to this. SanDisk once issued a press release about a photographer who shot pictures of the demolition of a bridge using a remotecontrolled camera. When the camera was struck by flying debris, the lens was shattered, but the flash card remained intact, yielding close-up photos of the explosion. Another incident involved the fiery crash of a jetliner whose flight recorder was severely burned. Even the epoxy packages on the flash chips were melted. A reverseengineering firm was able to reattach wires to the flash chip, and subsequently read out the data that enabled them to find the cause of the crash. How much abuse are these chips capable of withstanding? Standard specifications on a chip of any sort are operating temperatures of −40° to +125°C, with storage temperatures of up to 150°C. NAND flash chips follow this convention. The maximum storage temperature is set to one that the package can withstand for extended periods without losing its hermeticity, as moisture is lethal to a chip. The chip inside can withstand temperatures up to about 450°C before the silicon melts to the point that the impurities start to move around, changing the operation of the device. The epoxy will melt at a far lower temperature of over 300°C, but the chip will still keep its data after the package is long gone. Shock and vibration specifications on semiconductors are similar, in that they are specified conservatively to account for weaknesses in the package, rather than in the silicon chip itself. Typical specifications are for high levels of constant acceleration that is used to test if the bonding wires will pull off the chip, which once again is a test of the package rather than of the silicon itself.

4.15 Using Memory Reliability Specifications to Estimate Product Lifetime There are few specifications that must be considered to understand failure in a flash chip. This is particularly important to understand if the engineer wants to design a controller rather than to purchase one from a controller vendor. The most important specification is endurance. This is a measure of how many erase/ write cycles can be performed before a bit failure is likely to occur. The guaranteed minimum for most makers of SLC flash is 105 or 100,000 erase/write cycles. For

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

93

two-level MLC the guaranteed minimum is almost always 104 cycles, or one tenth the number of SLC. The next most important reliability specification is data retention, which is the manufacturer’s guarantee of how long bits will remain reliable if they are never rewritten in a worst-case environment. A time of 10 years is the norm here, which is usually longer than the life of the device that is using the flash. In order to really understand endurance, an engineer must know the erase and programming specifications of the device. Flash chips (NAND and NOR) are very sluggish to erase compared to other memory types, and erase cycles tend to be in the 1.5–2 ms range. These erases clear an entire sector or block. Each block is composed of pages, and a page programming time of 200 µs is typical. Other specifications required to understand these specifications are the block size (16 KB in a small-block NAND and 128 KB in a large-block NAND) and the page size (typically 512 B for a small-block NAND and 2 KB for a large-block NAND). Erases are performed on blocks, and writes are performed on pages. These and the overall size of the NAND chip will give the engineer enough to calculate a worst-case scenario for an endurance failure.

4.16 Flash Memory Cell Lifetimes and Wear Leveling Algorithms The real limiting factor that stands in the way of wear-out is the long time it takes to erase and reprogram a sector. A typical sector erase takes 1.5 ms and another 200 µs is needed to program a page within that sector. To perform these two functions 105 times would take about 23 minutes. This assumes that the system spends all of its time repeatedly erasing and programming a single sector in an attempt to cause it to fail. Let’s now look at a 4 Gb chip, which was a very typical size for an SLC NAND at the writing of this chapter. Such a device has 2,048 sectors. If a system spent all of its time erase/programming sectors in a round-robin order to ensure that all were hit equally, and if only one page per sector were programmed (the very worst case) a failure might occur in as soon as 34 days. More realistically, if every page were programmed in a sector before moving to the next sector, since this part has 64 pages per sector, the exercise would consume 258 days, or about 8½ months.

www.newnespress.com

94

Chapter 4

Remember that this example assumes that the device is never being read from, it is simply undergoing constant erase/writes. A few systems might behave this way, for example a black box recorder, but only if the bandwidth needs were rather extreme. Otherwise there would most likely be breaks between these cycles, prolonging the time before a failure. If a system spent a certain portion of its time idle, another portion reading, and another portion performing erase/writes, the 8½ months just calculated would most likely stretch out for a number of years. Also recall that the sole consequence of a bit failure will be that a block will be disabled, effectively shrinking the size of the chip. There will be no catastrophic failure. The NAND will not have to be replaced unless every last sector is needed. This is why wear leveling works so well. The difference between the first example’s 23 minutes and the 8½-month case attests to what can be accomplished if the controller is smart enough to track every sector’s activity. To look at this another way, wear leveling in a chip with 64 pages per sector and 2,048 sectors per chip turns the 105 erase/write cycles into an effective 1010 cycle, getting an improvement of one hundred thousand times. Note that this 1010 refers to the number of page writes that can be performed into the chip before a single bit error is likely to occur if the pages are being written into in sequence. Assuming that 1 MB photo files are being written into the chip this way (consuming 512 pages per photo), 256 million photos would have to be cycled through the part before a bit failure would be likely to occur. Wear leveling spreads out the problems of endurance or wear across the entire chip, so the bigger the chip the less frequently a wear failure will strike. With each doubling in a chip’s density the number of photos or the time between failures will also double. One company has posed this as an argument for decreasing endurance specifications. Since the incidence of bit failures halves for every doubling of chip density, does it make sense for a 4 Gb chip to have to exhibit the same endurance as its 256 Mb predecessor? Perhaps future chips will quietly relax this specification, counting on sheer size and wear leveling to hide this change from the user’s eyes. Assuming that the chip is jam-packed with information with no free space at all, each update will require that data be exchanged between a relatively inactive sector and a more active one, when the system tries to write into the more active sector. This means that each erase/write will result in two erase/writes to manage wear leveling. Although this could cut the endurance of the chip in half, this unlikely scenario still would hit a limit of 128 million photos or 4¼ months of nonstop writing.

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

95

4.17 Predicting NAND Bit Errors Based upon Worst-Case Usage The above analyses have been performed only for the extreme case. For real-life evaluations a more practical approach is in order. The designer needs to either understand usage patterns for actual analysis, or to make assumptions (which will need to be fully disclosed to the boss) that will provide a reasonable usage scenario. Let’s take an example of an MP3 player. We will assume that a player has 8 GB of storage and can store 2,000 songs. The most ambitious user might download an entire new collection of 2,000 songs once a day. If we choose a two level MLC NAND (typical) with a 104 endurance, and assume that no bit errors can be tolerated by the controller (unlikely) the simplest equation says that the MP3 player can be reprogrammed 104 times before a bit error is likely to be encountered. This means that it could fail after 104 days, or about 27 years. If a smaller number of songs is stored, and the controller performs wear leveling, hose 27 years would be extended by the inverse of the fraction of the memory that is used (i.e., if ¾ of the memory is used, then the endurance would extend to 4 ⁄3 times 27 years or about 36 years). Of course, this is all academic because a consumer electronics device is unlikely to be used for even five years before being replaced. Let’s evaluate another scenario. Although it is uncommon for an amateur photographer to spend the entire day shooting photos every day of the year, it is conceivable that a professional photographer would take one photo every five seconds for an 8-hour workday every workday of the year. This would amount to about 1.4 million photos per year. At this rate, if the photographer constantly used the same 512 MB card (one based on the 4 Gb chip used in the preceding analysis) to shoot 1 MB photos, then that card would have a bit error about once every two years. We should note that a 512 MB card is pretty small for today’s professionals, especially for the photographer in this example. At the five-second rate, the photographer would be shooting 720 photos per hour, filling the card relatively frequently, so time would have to be taken simply to read the photos off the card, or to erase them in the camera. It is most likely that this photographer would opt for a significantly larger card. Purists shoot in raw format, which might expand that 1 MB photo size used in the preceding analysis to about 45 MB. This would get the photographer only eleven photos on that card, or about 1 min of shooting.

www.newnespress.com

96

Chapter 4

As pointed out before, the frequency of failures declines as the card size increases, so this particular usage pattern would most likely require the photographer to suffer from a bit failure every several years. Both of these relatively extreme users see a far longer time between failures than the 8½ months outlined in the previous worst-case scenarios. This illustrates the need to understand the usage expected in the system to understand expected failure rates.

4.18 Flash Memory Format Specifications and Characteristics There are numerous flash card formats. This section will give a brief overview of the predominant types.

4.18.1 PCMCIA and Compact Flash (CF) These two cards offer the same electrical interface, but have different form factors. The PCMCIA card is the granddaddy of flash card formats, originally conceived as a means to allow PC users to add everything from DRAM to peripherals. The PCMCIA format is 85.6 × 54.0 mm in length and width, and supports three thicknesses, 3.3, 5.0, and 10.5 mm. The CF format is the same 3.3 mm thickness of the smallest of the PCMCIA cards, but is 36.3 × 43.8 mm in length and width. A 5.0 mm thickness is also specified, but this specification appears never to have been used. CF cards were designed to be limited to flash memory, and do not support peripherals and other memory types, although very small form factor hard disk drives are now available in the CF format. Both cards have a 50-pin socket (as opposed to the card-edge connectors used by all other card formats) and communicate by an ATA interface that supports capacities as high as 137 GB. Data transfer rates have been upgraded from 8 MBps to 66 MBps, and could be further enhanced in the future. The original PCMCIA card also can support a linear (non-ATA) interface.

4.18.2 SmartMedia and xD-Picture Card In an effort to reduce both size and cost, Toshiba and certain key camera manufacturers introduced the Solid State Floppy Disk Card or SSFDC, which was later given a simpler name of SmartMedia card. This was a thin plastic wafer, measuring a scant 0.76 mm thick, which used card-edge connectors and did not include a flash controller in the

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

97

card. The SmartMedia card’s full dimensions were 45 × 37 × 0.76 mm. This format’s maximum capacity was 128 MB, which was relatively large when the format was first introduced in 1995, but ran out of steam in 2002, when it was replaced with the xDPicture Card format. The xD (for Extreme Digital) card has a smaller length and width of 20 × 25 mm, but is thicker, at 1.7 mm, allowing multiple NAND chips to be stacked inside the card. The xD-Picture Card’s maximum capacity has been raised to 8 GB.

4.18.3 Multimedia Cards (MMC) SanDisk and Siemens introduced the MMC in 1997. Smaller than a SmartMedia card, this format is 32 × 24 × 1.4 mm and includes a controller. A card edge connector saves space and cost. This format has undergone a series of evolutionary changes, shrinking in size to support the MMCmobile format at 24 × 18 × 1.4 mm, then the MMCmicro at 12 × 14 × 1 mm. Another evolutionary change has been the communication speed. Read speed has been increased from 2.5 MBps to 52 MBps over time, while write speed has increased from 200 KB/ps to as high as 10 MBps, limited only by the speed of the NAND chips within the card. The maximum capacity supported by the MMC format is 4 GB.

4.18.4 Secure Digital (SD) Cards Shortly after the introduction of the MMC, the Secure Digital or SD card was introduced. This card has the same mechanical format as the MMC, and uses a somewhat compatible interface (MMC cards can be used in an SD slot, but not vice versa), but included a more sophisticated controller that addressed the security concerns of the media business who wanted some assurances that media would not be replicated with abandon by the users of these cards. As card formats shrank, so did the SD card, moving to the miniSD at 21.5 × 20 × 1.4 mm, followed by the microSD at 11 × 15 × 1 mm. The SD card interface tops out at a 12.5 MBps communication rate, and the format supports capacities of up to 32 GB. The MMC specification was upgraded after the introduction of the SD card to include digital rights management features.

4.18.5 Memory Stick Sony has its own format of memory card that the company calls the Memory Stick. This is also a card with an internal controller and a card-edge connector, and

www.newnespress.com

98

Chapter 4 Table 4.1: Key Attributes of Leading Flash Card Formats Format

Volume (cm3)

Maximum Capacity

Read Speed (MBps)

15.3 5.1 1.3 0.9 1.1 0.6 0.2 1.6 0.6 0.2 3.0 1.0 0.2

137 GB 137 GB 128 MB 8 GB 4 GB 4 GB 4 GB 32 GB 32 GB 32 GB 32 GB 32 GB 32 GB

132 132 20 5 52 52 52 12.5 12.5 12.5 19.7 19.7 19.7

PCMCIA CompactFlash SmartMedia xD-Picture Card MMC MMCmobile MMCmicro SD Card MiniSD MicroSD Memory Stick Memory Stick Duo Memory Stick Micro

measures 50 × 21.5 × 2.8 mm. As with the previously described formats, the Memory Stick has undergone a progression of shrinks to 31 × 20 × 1.6 mm (Memory Stick Duo), then 15 × 12.5 × 1.2 mm (Memory Stick Micro). The electrical interface has also progressed from an original 2.5 MBps read speed to the current 19.7 MBps read speed. The new specification is called Memory Stick PRO. The original specification was also limited in capacity to 128 MB, but the Memory Stick PRO specification will support capacities of up to 32 GB. The internal controller supports Sony’s proprietary digital rights management scheme through the Memory Stick’s internal controller. Table 4.1 compares the different card formats from a very high level of maximum bandwidth, capacity, and volume.

4.19 Flash Memory and Other Solid State Storage Technology Development 4.19.1 U3 and Applications Run from Flash Memory One approach to selling flash is to bundle self-launching applications inside the storage device itself. This is regularly seen in the world of CD-ROMs, where self-launching code is common.

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

99

Perhaps the most prevalent piece of self-launching code that is sold along with flash is the U3 bundle, a high-security desktop that a user can employ to attain a virtual desktop on any Windows PC around the world. The U3 concept is that a user will keep all their files in the USB drive. This lets the user feel free to use any PC anywhere. Everything from personal files to e-mail is stored in a U3 drive that will automatically synchronize with the user’s own computer whenever returned to that system. Should a host PC not have certain required software, a U3 drive will supply a reasonable facsimile of that application to allow the user to accomplish the necessary task. When a U3 drive is removed from a host system, all of that drive’s information will be deleted from the host, removing any security concerns.

4.19.2 Roadmap for Flash Memory Development The driving factor for flash card and USB flash drive capacity increases is the increase in capacity of NAND flash chips. NAND chips account for as much as 95 percent of the manufacturing cost of a flash card or USB flash drive. The increase in NAND chip capacity drives not only the amount of flash memory that can be fit into one of these card or USB formats, but it also drives the affordability of that amount of storage. At the writing of this book, a 2 GB USB flash drive is very affordable at under $20, whereas only one year earlier it would have cost twice that amount. Future readers are likely to laugh at how expensive these numbers seem. NAND chip vendors drive their costs down and their capacities up by shrinking the minimum size feature they can image onto a piece of silicon at a continuing progression. At the writing of this book, most NAND flash vendors produced their devices using processes with minimum feature sizes of 90 nm (nanometers, or millionths of millimeters) and 70 nm. These devices were most often manufactured using the MLC technique which allowed the capacity of the chip to be doubled through some hardware tricks. The cost of a chip is inversely proportional to the square of these numbers, that is, a 2 Gb (gigabit) chip produced on a 70 nm process would be 702/902 or 60 percent of the size of a 2 Gb chip produced on a 90 nm process. This implies that the cost of a 70 nm 2 Gb chip should be about 60 percent of that of its 90 nm counterpart. This also implies that certain applications that would have just been able to fit in a 2 Gb 90 nm chip may have enough room inside to accommodate a 4 Gb 70 nm chip. As the process shrinks, the capacity that can fit onto a card is likely to go up and the price for that much capacity has room to go down.

www.newnespress.com

100

Chapter 4

16M LOCOS LOCOS

~0.35um 0.35um

32M 1

SA-STI SA SA-STI

64M

Control Gate Tunnel Oxide

Control Gate Floating Gate Floating Gate ONO ONO WSi WSi

90nm New 90nm New Structure Structure 90nm~

0.25um~0.13um 0.25um 0.13um

ONO

WSi

Control Gate

Floating STI Gate

Floating Gate Tunnel Oxide

Tunnel TunnelOxide Oxide

LOCOS LOCOS

256M

Cell Size (mm2 )

Floating Gate Control Gate

STI

512M

0.1

Floating Gate

MLC

1G

1G

LOCOS

Control Gate

Floating Gate Control Gate

2G

2G 4G

STI

4G

8G

8G

0.01 STI

16G

16G

Multi Multi Level Level Cell Cell Fine Fine Lithography Lithography 0.001

250nm nm 250 250nm

160nm nm 160 160nm

130nm nm 130 130nm

90nm nm 70nm nm 90 70 90nm 70nm

32G

556nm 6nm

44xnm xnm

New New Materials Materials JanJan’93

JanJan’94

JanJan’95

JanJan’96

JanJan’97

JanJan’98

JanJan’99

JanJan’00

JanJan’01

JanJan’02

JanJan’03

JanJan’04

JanJan’05

JanJan’06

JanJan’07

JanJan’08

JanJan’09

Figure 4.10: Toshiba’s Process Migration for NAND Flash Memory.1 Figure 4.10 illustrates one vendor’s road map to shrinking processes. Toshiba, a leading NAND vendor and the inventor of both NOR and NAND flash technology, shows how they have progressed from 350 nm down to 56 nm, and how they plan to move beyond that to the 30 nm range. This represents 10-to-1 shrink in minimum features, yielding a 100-to-1 reduction in cost, and a similar increase in chip density. Other techniques (changes in cell structures in the orbs at the top and changes to MLC as delineated in the lower line of higher capacity products, have pushed those reductions even further.

4.19.3 Expected Growth in Storage Capacity for Flash Memory As Figure 4.10 illustrates, there is a steady progression in NAND chip capacity. This trend grows the amount of storage that can be sold at certain price points and how much storage can be fit into a given form factor. This evolution drives the chart in 1

Source: Toshiba Corporation, January 2007. Used with permission.

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies $10,000.0

$1,000.0

$100.0 1 $10.0

Price per Gigabyte

10

20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11

Average Density per NAND Chip

100

101

0.1 $1.0 Average Density Price per Gigabyte 0.01

$0.1

Figure 4.11: NAND Average Chip Density and Price per Gigabyte Trends.

Figure 4.11, which illustrates the likely average capacity of a NAND flash chip for the next five years, and the historical perspective for the past seven years. The chart uses logarithmic y axes, since the use of a linear y axis makes all but the extreme ends of the chart very difficult to read. Beyond 35 nm, the progression of NAND becomes less clear, since we do not yet know how to develop flash that will successfully operate past the 35 nm node. It is possible that the industry will move to a new approach to NAND (like Toshiba’s switch from LOCOS to SA-STI) but if flash cannot be coaxed past this point, then costs will stop their steady decline, making it likely that some alternative technology will be able to push past NAND’s cost structure to displace the current technology and take over the NAND market.

4.20 Expected Change in Cost per Gigabyte of Flash Memory Formats As was mentioned before, the major cost component in flash cards and USB flash drives (and any other flash-based storage medium) is the cost of the NAND flash chips. It is only reasonable to assume, then, that the costs of flash cards are likely to follow this in

www.newnespress.com

102

Chapter 4

Average Capacity of a Flash Card (GB)

100

10

1

0.1 $100 Card $20 Card

0.01

03

02

01

04 20 05 20 06 20 07 20 08 20 09 20 10 20 11

20

20

20

20

20

00

0.001

Figure 4.12: Growth Trends for $20 and $100 Flash Cards. some way. It is unlikely, though, that flash cards will go from $20 today to $10 next year, then $5 the following year, falling below $1 and onwards as the price for a gigabyte of flash falls. Instead, the price of cards is likely to follow certain consumer price points (say $20 and $100) and that the capacity of these cards will rise to meet these price points. Figure 4.12 is an illustration of how that is likely to happen. In this chart we have taken the prices used in Figure 4.11 to determine the average capacity of cards priced at $20 and $100 over the same time frame. Certain assumptions have been made in this chart about other costs and about retail markup, but the trend is clear, the average capacity of a $20 card is likely to double every year as costs decline by a similar rate.

4.21 Other Solid State Storage Technologies There was some mention above about the difficulties faced by flash designers as manufacturing processes continue to shrink. Since manufacturers have been anticipating a roadblock to continued process migration for quite a while now, they have invested in alternative technologies. The prevailing technologies are ferroelectrics, magnetic memories, and phase-change memories. Each of these has an acronym: FRAM, MRAM, and PRAM, in that order. There are several other contenders, including carbon

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

103

nanotubes and silicon nanocrystals, but these technologies do not have the funding of the big three. Each of these technologies has certain strong advantages. All three offer faster write cycles than NAND, and all but FRAM boast unlimited wear. Furthermore, they are all low-power technologies that promise to scale to ever-tightening processes better than is expected of NAND after 35 nm. Are these a threat to NAND? It is likely that one of these technologies could take over the NAND business once NAND has hit a brick wall and cannot be shrunk any further. Given that this inevitable brick wall has been pushed out recently by two process generations, it appears quite possible that there will continue to be life in flash after 35 nm, and perhaps for several more process generations. In other words, as with hard disk drives and other storage technologies, brick walls are cast in sand. Cost is what drives acceptance in the world of semiconductor memory, and since these materials are less well understood than silicon, the only way they can catch up to the semiconductor memory cost structure is if semiconductor memory stops in its progress, giving these other technologies a chance to catch up and surpass NAND’s per-gigabyte costs. It is reasonable to say that NAND will be replaced by one of these technologies before this book disintegrates from age, but it does not appear reasonable to assume that they will gain much ground over the next five years. At the writing of this chapter, two leading flash vendors, Samsung and Intel, have cast their votes with PRAM, which they believe to be the technology with the highest chance of outstripping silicon’s cost declines over the next several years.

4.22 Chapter Summary •

Flash memory technology is an application of floating gate semiconductor technology. Flash memory is a nonvolatile storage technology that is widely used in mobile consumer electronics applications.

•

The basic operations of a flash memory device are reading, writing, and erasing. In order to rewrite data on a flash memory device, the flash memory cells must first undergo an erasure step.

•

Over repeated use, the oxide tunneling layer in a flash memory device can accumulate charges from electrons that do not tunnel through. As the number of trapped charges increases, it becomes difficult to keep a charge in the floating

www.newnespress.com

104

Chapter 4 gate. This accumulation of charge is referred to as wear of the flash memory cell, and eventually leads to the inability to maintain data on that cell.

•

There are two kinds of flash memory, NOR and NAND. The two terms are names of types of logic gates, the negated OR function and the negated AND function. The difference between the two types of architecture is real estate. NAND has a significantly smaller die size than does NOR. This translates to significant cost savings. NAND is different from other memory technologies in that it is partially serial and partially parallel.

•

The serial technology in NAND is very prone to bit errors. Like hard disk drives, NAND makers used error correction technology to correct bit errors. A sector of data will have some parity bits added to check for errors in the sector data. These check or parity bits are programmed with a highly compressed code that indicates what the data in those other 256 bits should look like.

•

Wear leveling is a technique to spread the writing and erasing over all the cells in the flash memory so that individual cells do not wear out before the rest of the memory. Wear leveling is managed by the flash memory controller chip.

•

The controller is also responsible for bad block management in the memory. The controller keeps track of the number of bad blocks in each sector, and if the number of bad blocks becomes too large, the sector is declared bad and not used anymore. The data that was on that sector is then moved to a spare unused sector.

•

Flash memory can be used as a chip installed within the circuit board of a device, or it may be embedded in a removable card.

•

Like a hard disk drive, the flash contains files, and those files must be managed. A flash file system makes the data on the flash memory device look as if it comes from a hard disk drive to the host device.

•

In order to increase the storage capacity for a NAND flash memory device with a fixed number of memory cells, the number of bits per cell can be increased. Increasing the number of bits per cell involves splitting the voltage levels that are used to read the bits on the cell. Multilevel cells (MLCs) trade off increasing storage capacity for greater sensitivity to flash wear during the erase and write cycles. Also, the noise on the cells does not decrease as the cell voltages are split, and so there are more errors that need correction with an MLC flash

www.newnespress.com

Fundamentals of Flash Memory and Other Solid State Memory Technologies

105

memory. Implementation of MLC involves more sophisticated controller technology, and this sophistication and processing power has to increase as the number of bits per cell increases. •

Flash memory dies can be stacked on top of each other to achieve greater volumetric storage capacity. Flash memories are particularly good at this, since they generate lower operating power (that needs to be dissipated) compared to other types of semiconductor chips.

•

Flash reliability is dependent upon careful control of cell wear and error correction. Since these operations occur in the flash memory controller, controller technology is key to providing specified reliability. This is even more important with MLC flash. Flash memory is inherently insensitive to shock or vibration, since it is a solid state storage technology. Used in the right way and in the right applications, flash memory can give many years of useful life in a consumer product.

•

Removable flash memory card formats have developed over the years due to new applications, form factors, and proprietary designs. Today, the most common flash card formats include MMC, SSD, memory card, and miniaturized versions of these formats.

•

Over the last few years, flash memory has undergone a significant reduction in price for given storage capacity points. This price reduction expressed in $/GB should continue at about a 40 percent annual rate for the near future. As affordable flash memory storage capacity increases, the number of applications that flash memory can participate in increases. With the lower base price for flash memory (compared to other technologies such as hard disk drives) flash memory will be found in lower storage capacity consumer electronics products with growth potential into higher capacity products.

www.newnespress.com

This page intentionally left blank

CHAPTE R 5

Storage in Home Consumer Electronics Devices Objectives in This Chapter •

Look at the functions, design trade-offs, and uses of digital storage in common household consumer electronics products.

•

Consider the development of digital storage in digital video recorders and how the desire of consumers to maintain access to content they record could lead to direct attached as well as network attached additional digital storage in the home.

•

Compare in-home network storage for DVRs to head-end storage initiatives to consolidate digital storage needs, particularly for service providers.

•

Examine how home storage networking could expand to create home media centers using appliance devices and home media adapters with network storage of multiple-purpose products such as game systems.

•

Look at video requirements in the future and their impact on digital storage capacity requirements in the home.

5.1 Introduction Digital storage is an enabling technology for many modern consumer devices. This includes products such as Digital Video Recorders (DVRs), home media centers, home network storage, automobile entertainment and navigation systems, mobile music and video players, digital still cameras, camcorders, cell phones, game systems, and many other devices. Digital storage allows the use of rich content in these devices as well as

www.newnespress.com

108

Chapter 5

allowing personal content creation. This chapter and the next explore many of these devices. It shows how digital storage is used, the important characteristics for digital storage used in these applications, and finally where each of these applications could be heading in the future. We will pay attention to how digital storage will impact the development of these devices.

5.2 Personal Video Recorders and Digital Video Recorders Digital video recorders, also known as Personal Video Recorders (PVRs) have dramatically changed customer viewing habits. A digital video recorder can record and play back full motion video and audio signals using the ISO MPEG audio/video standards for the compression of audio and video content. This differs from the analog Video Cassette Recorders (VCRs) that were the first commercial devices to free consumers from a fixed television program schedule. Storage in digital video recorders is usually done with hard disk drives. Ampex, however, introduced a hard disk analog video recorder in 1967, the HS-100. The HS-100 was developed to give CBS instant freeze-frame capability. The product recorded composite analog video onto a 14-inch diameter hard disk using FM modulation. The total storage time was only 30 seconds, but the device could record continuously and play back from twice normal speed down to still frame. In 1985, an employee of Honeywell’s Physical Sciences Center, David Rafner, first described a drive-based DVR designed for home TV recording, time-slipping, and commercial skipping. Rafner’s idea (U.S. patent 4,972,396) focused on a multi-channel design to allow simultaneous independent recording and playback. Broadly anticipating future DVR developments, it describes possible applications such as streaming compression, editing, captioning, multi-channel security monitoring, military sensor platforms, and remotely piloted vehicles.1 TiVo and ReplayTV launched their digital video recorder products at the 1999 Consumer Electronics Show. Since their introduction, DVRs have revolutionized home video viewing habits in a way that VHS-based home video recorders could not. By using digital recording technologies, it became easier to access particular content and store it on hard disk drives. With the introduction of electronic program guides for these products, they could be set to automatically obtain video content from a cable or satellite network source and store it for later viewing. This intelligent time shifting has forever 1

Wikopedia article on digital video recorders, http://en.wikipedia.org/wiki/Digital_video_recorder.

www.newnespress.com

Storage in Home Consumer Electronics Devices

109

Figure 5.1: TiVo Digital Video Recorder.

changed consumer viewing habits and threatened traditional advertising models, since customers have been able to skip ahead or fast forward through undesired advertising. Besides stand-alone DVRs produced primarily by TiVo in the U.S. (and vendors such as Thomson, Topfield, Fusion, Pace, and Humax elsewhere in the world) this technology is often embedded in modern cable or satellite set-top boxes where such features have helped prevent customer churn. They are also often bundled into DVD recorders that can be used for archiving video content. This last feature is particularly popular in Asia. Companies that manufacture set-top boxes (STBs) with DVR capability include Cisco’s Scientific Atlanta, Motorola, and Moxi. DVR capability has also been incorporated into many Home Media Center PCs and even into some TVs by companies such as LG and Hitachi. Video recording capability appears to be very popular and it is likely that this function will become common even on mobile devices.

5.2.1 Basic Layout and Design of Digital Video Recorders Figures 5.1 through 5.3 show a DVR product manufactured by TiVo (one of the founders of this business).2 Figure 5.1 shows an example of digital video recorder set to record up to 80 hours of standard definition TV. This product came with an 80 GB internal hard disk drive manufactured by Maxtor (now part of Seagate Technology). Series 2 TiVo players like these include a USB port (for Ethernet and 802.11 adapters), a faster CPU than the original units, and more RAM. TiVo takes advantage of this network connectivity with new software features like TiVo ToGo and Home Media Engine 2

Inside Peek report on the TiVo Digital Video Recorder, 2007, http://www.inside-peek.com/.

www.newnespress.com

110

Chapter 5

Figure 5.2: Interior of TiVo Digital Video Recorder. applications. Figure 5.2 shows the interior of the TiVo box. Hard disk drives with a 3.5-inch size are used in these products, since they provide the highest capacity at the lowest price. As can be seen, there is a lot of empty space in the box. Later, we will discuss how DVR storage and products may develop and could become more compact and less expensive in the future. Figure 5.3 shows a close-up of the main circuit board of this device, identifying several of the Integrated Circuits (ICs) on the board. Figure 5.4 shows a reference design for a digital video recorder from Texas Instruments. The DVR block diagram includes several components: •

The Stereo Audio Coder/Decoder (CODEC) uses audio analog to digital conversion and digital to analog conversion to digitize and playback analog audio.

•

The Triple Digital to Analog Converter (DAC) converts digital video into analog video output in various supported formats such as NTSC/PAL, S-Video, and YPrPb component video. High performance op-amps are required in the output stages to amplify the video signals.

•

The Analog Video Decoder turns analog input signals into a digital form and decodes the baseband analog formats (NTSC/PAL/SECAM) into digital component video.

•

The MPEG Video Encoder/Decoder and Audio Processor does the MPEG video encoding/decoding and multiplexing/demultiplexing of the audio and video bit streams.

www.newnespress.com

Storage in Home Consumer Electronics Devices Atmel AT90SC646 4C 0628 622532

111

Etron EM638325TS-6G K183705ACC637-13M

Broadcom Kfir-II BCM7040KQL TN0617 P10 779939

MK2745-27R CL670105Z 0626 Micronas AVF4913A C2 6227I 31 Z3 F 206783.102

Nanya 0633 NT5DS16M16CS-5T 62120300AP SG

Broadcom BCM7317KPB1 TA0617 P11 780687 P2

SST 37VF010 70-3C-NH 06I6065-A

f1G22 MJD 45H11

M37252 BR 0534 MAL

Si0318-FS 0609CG-LAN2 Si2434-FT 0612CCAAP PH

Figure 5.3: Main Circuit Board of TiVo DVR Showing Many of the ICs.

•

The Central Processing Unit (CPU) controls the DVR operating system software, overlay or text/graphics, and the user interface.

•

The ADSL cable modem allows the user to control the DVR using the panel keypad or the remote control for commercial units and for consumer use allows access to the electronic programming guide.

www.newnespress.com

112

Chapter 5 Stereo Headphone

Microphone L Stereo Audio Line In

L Stereo Video CODEC

R

Stereo Audio Line Out

R

Triple DAC

Analog Video In Analog Video Decoder

Clock Sources

SDRAM

MPEG Video Encoder/ Decoder Audio Processor

Color Space Conversion Video Encoder

FLASH

RGB Out TV CH3-4 MOD

RF Video Out

Triple DAC

S-Video Composite Video

Triple DAC

Component Video

OSD ROM SDRAM

Plug ADSL Cable MODEM

Storage Media Ethernet PHY Transceiver

AC Adapter

Host CPU

USER INTERFACE

FLASH LCD Display

RS232/ 442 FPGA PCI/Bridge

Micro Controller

RTC

Keypad

Remote Control

LAN Port

Main Supply

LED Supply

LCD Supply

DC/DC Buck Converter

DC/DC Boost Converter

DC/DC Boost Converter

DSP/CPU Core Analog Supply Supply DC/DC Buck Converter

LDO

LEGEND

Digital I/O Memory LDO

Processor

Supply Voltage Supervisor

Integrated Power Management

Interface Amplifier Logic Power ADC/DAC Other

Figure 5.4: Reference Design Block Diagram for Digital Video Recorder.3

3

•

The Floating Point Gate Array (FPGA)/Peripheral Component Interconnect (PCI) interface provides command and data transfer between devices using the PCI bus.

•

The user interface allows the user to control the DVR using the panel keypad or the remote control.

•

The AC adapter and the DC/DC Buck Converter brings the input power into the box and converts it to the voltage levels required for the various functional blocks.

Courtesy of Texas Instruments.

www.newnespress.com

Storage in Home Consumer Electronics Devices

113

5.2.2 Digital Video Storage Requirements and DVR Storage Design Richer video content requires higher capacity storage as well as faster data rates. For instance, a standard definition video requires about 1.25 GB per hour of storage while an HD video requires about 8.3 GB per hour. Thus a 500 GB disk drive can hold about 400 hours of standard definition and only 60 hours of HD content. An HD recorder then requires access to larger data storage. This requires the use of larger capacity hard disk drives with these products. Such greater storage capacity could be provided internally with an embedded hard disk drive, through a Direct Attached Storage device (DAS or non-shared storage) on the back of the DVR or through a network connection on the DVR attaching the device to network attached storage. In addition to storage requirements, the data rate required by the device for storage and any network and external interfaces depends upon the resolution of the content and the desired special or trick features. The vast majority of broadcast content in the U.S. today is MPEG2. Measurements of streams from the top four U.S. service providers show that SDTV MPEG2 broadcasts are variable bit rate with average data rates between 2 and 3 Mbps, while peak rates are around 9 Mbps. Video On Demand (VOD) SDTV streams tend to be a constant bit rate between 4 and 5 Mbps. HDTV MPEG2 streams can also carry the requirements of content providers for a minimum of approximately 12 Mbps allocation.4 Fast forward, reverse, and other trick modes increase the peak data rate by a factor of three or more if continuous streaming video is desired during the trick mode. Let’s take a look at the external storage add-on options to understand how such products are being designed and attached to DVR devices and how this approach can be used to avoid field replacement and upgrades of DVR in set-top box internal storage and to reduce the upfront capital costs of DVR STB implementation.

5.2.3 External Direct Attached Storage for DVRs The earlier generation of digital video recorders did not have any active options to expand digital storage capacity beyond that built inside the DVR. As a consequence, users have had to erase programming that they have stored in order to record new programming. Set-top box manufacturers have incorporated DVR capability to create increased customer loyalty and are now the biggest suppliers of DVRs to the world. In 4

Leveraging MOCA for In Home Networking, Anton Monk, http://www.convergedigest.com/bp-ttp/bp3. asp?ID=453&ctgy=Home.

www.newnespress.com

114

Chapter 5 Table 5.1: Comparison of Internal and External Storage Interfaces eSATA (External)

SATA 300 (Internal)

PATA 133 (Internal)

FireWire 800 (External)

FireWire 400 (External)

USB 2.0 (External)

Ultra-320 SCSI (External)

3,000

3,000

1,064

786

400

2,560

2

1

0.46

Power Provided

No

No

No

Devices per Channel

1 (15 with Port Multiplier)

1 per Line

2

4.5 (16 Cables Can be Daisy Chained up to 72 m) Yes (12– 25 V, 15 W) 63

4.5 (16 Cables Can be Daisy Chained up to 72 m) Yes (12– 25 V, 15 W) 63

480 (burst) 5 (USB Hubs Can be Daisy Chained up to 25 m) Yes (5 V, 2.5 W)

Data Rate (Mbps) Maximum Cable Length (m)

127

12

No

16

order to control their costs for equipment in the field, STB manufacturers want to ship product using the current volume disk drive rather than higher capacity drives that cost considerably more. Thus most DVRs do not have very high capacity hard disk drives, nor is there much incentive to supply higher capacity disk drives in STB DVRs. Table 5.1 shows technical comparisons between various external and internal storage interfaces showing differences in data rate, cable length, power on the line, and devices per channel. The fastest interfaces are the SATA and eSATA interfaces as well as the SCSI interface. The SCSI interface is used on higher performance (and more expensive) storage systems and so is not common on consumer products. Common interfaces in consumer products are USB and Firewire (IEEE 1394). Firewire interfaces have been primarily used as a video recorder connection interface to move content from a camera to a video editing workstation. eSATA is a new interface technology based upon the serial ATA (SATA) internal disk drive interface that allows very fast data transfers (3 Gbps) and also multiple device connection using port multiplication. SATA interface disk drives have been available since 2005, but 2007 marked the year that they became predominant over the older parallel ATA interface. By 2010 there will be no parallel ATA interface hard disk drives produced. The SATA road map calls for data rates reaching 600 MBps (currently SATA II is 300 MBps) probably before the end

www.newnespress.com

Storage in Home Consumer Electronics Devices

115

Figure 5.5: USB Expansion of a DVR Allows Only a Single Storage Device.

Figure 5.6: eSATA Expansion of a DVR Allows Many External Storage Devices.

of the decade. This very high data rate could also be extended to an external eSATA version as well. Consumers would like to have a greater capability to retain their DVR content and an external drive expansion option would be very attractive to many DVR customers. A USB port would allow adding a single expansion storage device to a DVR as shown in Figure 5.5, but a single eSATA port would allow additional expansion of a DVR storage capacity due to port multiplication as shown in Figure 5.6. Figure 5.7 shows how port multiplication works. Realizing that expansion of storage on DVRs is important to consumers, set-top box manufacturers such as Scientific Atlanta, Pace, and others have incorporated eSATA ports on their new STB offerings. This way, cable and satellite companies can offer STBs with DVR capability but not a very large hard disk drive inside the DVR itself. Customers can purchase their own eSATA external storage devices that they can use to expand the capacity of their DVRs. If the eSATA chip on the STB including DVR functionality includes drive locking (another feature on some eSATA electronics), then the external drives could only be used on the DVR. Figure 5.8 shows a Scientific Atlanta STB with an eSATA port. Figure 5.9 shows a Maxtor/Seagate eSATA external storage box targeted to the DVR expansion market. Other companies such as Western Digital also offer such eSATA

www.newnespress.com

116

Chapter 5 SATA SATA

HDD

F:

HDD

G:

PM

Figure 5.7: Use of eSATA Port Multiplication to Connect Multiple Storage Devices to a Single SATA Host Connection.

Figure 5.8: Images of Set-Top Boxes with DVR Capability and eSATA Ports for Storage Expansion.

Figure 5.9: Maxtor (Seagate) Quickview Expanders for Adding Storage Capacity to DVRs.

www.newnespress.com

Storage in Home Consumer Electronics Devices

117

expansion storage devices. Based upon the growth in DVRs and the issue with limited storage capacity, it is expected that DVR direct connected expansion external storage devices will be a very hot market in the next few years.

5.2.4 Network Attached Storage for DVRs The other option to increase the capacity of a DVR or set-top box with DVR capability is to provide a means to connect the DVR to a network. Such networking can be used for sharing recorded content with televisions, PCs, and other play out devices attached to the network or used to allow storage of video content on a storage device attached to the network, so that content can be saved even if the internal DVR storage is erased. The TiVo DVR described at the beginning of this section provided a USB interface that could be connected to an adapter that interfaces with a home network. Some set-top box DVRs made by Scientific Atlanta and other manufacturers have built-in Ethernet connections on the back of the box. This allows direct connection of the DVR and the network. Ethernet wired or wireless connections within the home can also be combined with internet connections to allow access of a DVR to content through the internet rather than just through on-the-air broadcast, satellite, or cable connections. Such a device combines the DVR with a device referred to as IPTV (after the internet protocol or IP used for internet traffic management) that records content from the internet for later play out. Apart from passive backup storage of content, such a network attached device could be part of a home media center. This will be covered in another section of this chapter. A variation of the network attached storage architecture is to have a virtual drive hosted at the cable head-end. Some cable companies are exploring this option. This allows the customer to scale their storage capacity through a monthly charge rather than by purchasing and connecting new hardware in the home. This drive could be used for programming or for PC data storage applications. It also gives the service provider a leg-up on providing IPTV services because they are in a better position to control the integrity of the source programming and the Quality of Service (QoS).

5.2.5 New Digital Video Recording Developments DVR products free individuals to deal with video and television on their own time rather than the timetable of the networks and other content providers. It also gives them

www.newnespress.com

118

Chapter 5

more control of the viewing experience, for instance avoiding or fast forwarding through ads. This makes this content more useful to the consumer. What are some other developments that could build on this useful functionality that might extend this benefit and reduce the costs and increase the availability of these devices? We will explore these potential developments here. Consumers often would rather not erase content they have recorded to make room for new content. This gives them more control over their viewing experience since they can watch an older show whenever they wish and they can build their own library of content. Also, the move to higher resolution video content, for instance the move from standard definition video to HD video, requires many times more storage capacity in order to hold an equivalent hour of content. A storage capacity of several terabytes (TB) to hold recorded TV content is not out of the question, and will probably be rather common within a few years. This will lead to the development of external storage solutions to expand the available storage of the DVR. Direct attached and network storage options for increasing DVR storage capacity have been explored earlier in this chapter. Again, another option is to have such huge quantities of storage located at the cable head-end and paid for as needed. Basic DVR capability appears to be generally in the public domain but use of DVR functionality with an electronic program guide is covered by current patents. TiVo and Replay TV are the primary patent and IP holders for this combination of technology. DVR capability built into modern electronics would usually include some license fees in the price of the product. Building such functionality into general electronics makes integration of DVRs into DVD players and other devices much easier and less expensive. Greater integration of DVR functionality will also allow such devices to be smaller since the circuit board can be much simpler and more compact. In Chapter 7 we will discuss further methods for integrating consumer electronics functions into the storage devices for the ultimate consolidation of electronics. As hard disk drive storage capacity increases, 2.5-inch hard disk drives will be available in the near future with several hundreds of GB of storage capacity (as of this writing they are higher than 200 GB in storage capacity). Using a smaller hard drive will help make smaller DVRs and devices that include DVR functionality and in particular a 2.5-inch drive based product could be made thinner than a DVR using a 3.5-inch disk drive.

www.newnespress.com

Storage in Home Consumer Electronics Devices

119

As time goes on, DVR functionality will also be combined with network connections and recording of streaming or downloaded content from the internet using technology such as IPTV. As home video recording devices become capable of recording content from a very diverse set of sources, electronic program guides will have to evolve as well, becoming more like an ongoing search agent that looks across a diverse set of sources to find the content that a customer is interested in. Hard disk drives are the recording devices used in most digital video recorders, but some DVR type devices could be made using other storage media. SanDisk introduced a product concept that they called TakeTV. They are working with television manufacturers to include a special USB interface on televisions that would allow plugging in a special USB based flash memory drive that could record television content, and that could then be plugged into a PC to play. As the price of flash memory declines, such a device could be able to record a fair amount of video content and allow transferring such content using a “sneaker net.” There are several mostly flash memory based portable music players with AM and/or FM radio reception that allow an audio version of a DVR for radio content. Combining such functionality with internet radio stations and podcasts as well as an appropriate content search capability could provide consumers with an enormous universe of audio content. While automotive satellite radio is currently enjoying much popularity, its monthly fee business model could be quickly destroyed by solid state or small form factor hard disk drive storage embedded in the car that downloads hours of fresh podcast content daily over Wi-Fi while parked in the garage and listened to throughout the day. HDTV is only now becoming commonplace, but researchers are already working on even higher definition television standards. Researchers in Japan are working on the Ultra High Definition TV initiative for next generation TV. Such next generation TV will require even greater storage capacity per hour than HD (16X HDTV resolution is envisioned for this future standard). The estimated time for this product to show up in homes is 20 years (2027). This will require even larger hard disk drives both within and outside the DVR box. Table 5.2 is an estimate of the storage capacity (internal and external) that a typical DVD enabled device requires over time. Of course the far-out data shows what could happen. Many things will change over that period of time, and even the name DVR may be an antique by then or taken for granted. DVRs have become a standard application that is increasingly combined with many other applications and has served as an inspiration to creating other devices for recording content and enjoying it on the customer’s own schedule and using many

www.newnespress.com

120

Chapter 5 Table 5.2: DVR Storage Requirements Over Time (Combination of Internal and External Storage)

Year

Internal Storage (GB)

External Storage

2006 2010 2014

40 80 160

0 1 TB 10 TB

2018 2024

320 640

100 TB 1 PB

Comments No Valid External Storage Options External Storage Options Available Assumes Ability to Retain Considerable Recorded Programming Lots of Stuff—Some Noncommercial Huge Capacity Anticipated

different recording devices. Since resolution requirements are increasing, these devices are becoming more popular. For this reason and because customers tend to keep content that they have recorded, the storage capacity required for these applications are increasing enormously.

5.3 Home Media Center and Home Network Storage A media center is a computer adapted for recording or playing content such as music, videos, photographs out of a local hard disk drive or other storage device, or over a network that is connected to some storage device where the content resides. These devices are often also able to play content from CDs or DVDs or other optical or other distribution media. Like many other entertainment devices in the home, they also usually have a remote control that can control the function of the device and the content can be viewed on a TV set. The media center can be a device designed for this function or a regular computer (such as a PC or game machine) with appropriate software. Media centers usually have a DVR function built into them to record video on hard disk drives. Media centers are also a natural venue for IPTV, where video and other content is taken off the web. Media centers can be created as single appliances providing network connectivity as well as content storage, or created using separate network storage devices and digital media adapter products. Dedicated digital media adapter products are produced by many networking and computer hardware companies. In addition, many other multiple-use devices such as game systems (e.g., the Xbox 360) can also be used as digital media adapters. For the sake of simplicity, we will refer to a single media center appliance, separate storage and media adapter combinations, as well as multiple use devices with

www.newnespress.com

Storage in Home Consumer Electronics Devices

121

storage (such as home game systems) that can act as media adapters, as media centers in this section. Applications where media centers have achieved some market demand as of this writing include: digital sign systems; automotive entertainment systems for limousines, tour buses, and car stereos; in-flight or room entertainment and video on demand systems; and mobile collections for travelers and audio/video collectors. Some of the software products that can be used to transform a PC into a media center are: •

Freevo (Linux).

•

Front Row (Apple).

•

GeeXboX (Linux).

•

Media Portal.

•

MythTV (Linux).

•

LinuxMCE (Linux).

•

My Media System (Linux).

•

SaveTV.

•

Tvedia.

•

Windows Media Center (Windows Vista and Windows XP Media Center Edition).

•

Xbox Media Center (not the same as Windows XP Media Center eXtender).

•

Elisa Media Center (Linux).

5.3.1 Basic Layout of Media Center Devices Various commercial devices are available for home and commercial markets offering media center capability. Figure 5.10 shows a block diagram of a Linux-based media center. A media center must provide the performance necessary for the deployment of functions like MPEG2/4

www.newnespress.com

122

Chapter 5

PowerPC MPX Bus

DIMM Flash PHY

DDR/DDR2 Local bus GigE

PCI/ PCI-X

IEEE 1394 South Bridge

System Controller

IDE/SATA HDD USB 2.0 IrDA

Slot TV Module Expansion Graphics TV/Video

Figure 5.10: Linux-Based Media Center Using PowerPC Architecture.

decode and encode, and DVR combined with electronic programming guide, gaming, internet browsing, and digital content editing. In addition, the media center may also act as a media server to clients throughout the home, as well as providing voice and video conferencing. In order to be deployed in a home environment, the system design must be streamlined to blend into the home environment. Designers of media centers must integrate a number of connectivity technologies, high performance processing capabilities (using embedded microprocessors), and a number of digital media (video and audio) processing technologies. Because of the complexity and emerging standards that may change with time, a significant portion of the system design is implemented in software. Some basic design features can significantly affect media centers and their general usefulness. It is important to keep these features and options in mind when designing these products. 1.

Will the media center try to connect various devices in the home, including computers and consumer electronics products?

2.

Will the media center connect to other networks in the home such as coaxial cable, phone lines, or power lines?

www.newnespress.com

Storage in Home Consumer Electronics Devices

123

3.

How well is the device able to handle commercial as well as personal content?

4.

Will the storage of content be inside the media center or on the same network as the media center?

5.

Can the media center play content located in other devices, and if so, which other devices can share this content?

6.

How will commercial and personal content be protected and how can privacy be ensured?

We will deal with these questions in the following sections.

5.3.2 Home Networking Requirements for Media Centers A media center is basically a computer with electronics and software designed to support multiple streams of content in a home environment. It can look like a personal computer or some sort of custom-built device designed for the home environment. The digital storage requirements depend upon the amount of content that they will contain. Since this could include both personal and commercial content, the actual storage required to be connected to the network could be very large. In addition, media centers are likely to take over the function of home DVRs and many other home consumer electronics devices. Thus the storage requirements for DVRs given in Table 5.3 also may be true for home media centers with additional storage capacity required for personal content. Since in my estimation personal content storage in a typical home will exceed commercial content by 2010, the storage estimates in Table 5.3 could be considerably higher, perhaps two to three times those mentioned in the table.

Table 5.3: Video Resolution, Single Stream Rate, and Storage Capacity Required per Hour for a Stream Video Resolution SDTV HDTV Ultra HDTV

Stream Rate (Mbps)

Storage Capacity/Hour (GB)

∼8.0 ∼19.3 ∼295

∼2.00 ∼8.89 ∼133

www.newnespress.com

124

Chapter 5

5.3.3 Home Media Centers and the Internet It is natural that a home media center network should also connect to the internet. Connecting to the internet will allow the home media center to access commercial content through services such as IPTV and also personal content that other people post on the web to share. Content sharing on the internet is one of the fastest growing uses and one that consumes quite a bit of bandwidth and storage in the data centers that handle content sharing. Security of commercial content and protection of personal content will be very important in order for such broad networking to take place. There will be some mix of downloading and uploading of content from physical storage media that could occur on media centers if content owners become comfortable with this concept. For personal content, downloading and uploading will be at the discretion of the user. A universal media center must be designed to access and use content whatever the source is. IPTV describes a technology allowing a home entertainment system to access video and other content through the internet, downloading and viewing content when it is desired. IPTV has time-shifting approaches similar to DVRs, and can result in significant digital storage requirements on the media center as content storage builds up. Like an STB it is likely that IPTV will result in external storage demand for longer term retention of recorded content. Figure 5.11 illustrates the service provider

In-Home Device STBs

Video Source Local Video Source National/Regional Video Source

DSLAMs Firewall

Firewall

Routers

Swithches

B-RAS Ethernet Aggregation Network

IP/MPLS Core

Network Infrastructure

Figure 5.11: Service Provider Infrastructure to Support IPTV.

www.newnespress.com

Storage in Home Consumer Electronics Devices

125

infrastructure to support IPTV. An IPTV receiver can be a PC, a special set-top box with Internet connections, or a monitor or TV with IPTV reception capability and an internet connection. An IPTV set-top box or a home media center incorporating IPTV may also support other functions enabled by internet connectivity. An important feature for internet connections is that data can be transported in both directions allowing for more built-in interactivity options than is typical in coaxial or satellite set-top boxes. This interactivity can be used to make an IPTV device capable of other internet functions. Since IPTV and other internet content access methods appear to be very popular and could soon be larger than traditional content distribution technologies, especially when sharing of personal content in also included, a centralized home repository for content such as a home media center should include these capabilities of gathering content in addition to cable and satellite. As time goes on, all these connections and networks could merge in some fashion, providing these features in a seamless service offering. That is, it will become a ubiquitous consumer offering if it can be done simply and is easy to use for consumers—this may be the biggest challenge and could take some time to perfect.

5.3.4 Future Media Center Capability The major factors driving media center capability over the next few years are: •

Higher resolution content requiring greater storage capacity and internal bandwidth requirements.

•

DRM and privacy protection of personal content.

•

Indexing and organizing content to make it easier to find and use material.

•

Unified backup of content on a media center, including deduplication of similar files.

•

Use of the internet to back up personal content stored on a home media center to prevent its destruction in case a disaster destroys the media center.

We will explore these factors below.

www.newnespress.com

126

Chapter 5 Table 5.4: Bandwidth Requirements for 16 Channel Media Center as a Function of Video Resolution Video Resolution

16 Stream Rate (Mbps)

SDTV HDTV Ultra HDTV

∼128 ∼309 ∼4,720

High Resolution Content for Home Media Centers Table 5.4 compares data rate and required storage capacity per hour of content for standard definition TV, HDTV, and a hypothetical next generation HD resolution. Note that actual numbers may vary depending upon the compression used. As content moves from standard definition to HD and eventually to even higher resolution, the bandwidth and storage capacities soar. A media center must be able to handle multiple streams of content in order to serve its function of sharing content and also must support trick modes such as fast forward and reverse. These trick modes increase the required bandwidth of the system by a factor of three or more if continuous streaming must be supported at the same time. In practice, media centers are being designed with the equivalent of 16 separate channels of content. An Ultra HDTV product with 16 streams would be an awesome device with total bandwidth requirements of 590 MBps. The network required for such a system would need to be equivalent to a 10 Gb Ethernet network. Such high bandwidths will not be commonly available for at least 10 years. Faster Organization and Finding of Content in Home Media Centers As we end up with potentially vast amounts of stored content in our home environments (both commercial and personal content in a home could reach petabyte levels before 2020) we will increasingly need some way to find this content. There will be three keys to the efficient access of potentially vast stores of home content: 1.

Efficient metadata creation to describe the content to a search engine.

2.

Effective indexing and context placing of content to speed up searches.

3.

A way to display the results of a search that allows the user the fastest means to find the content closest to what he or she was searching for.

www.newnespress.com

Storage in Home Consumer Electronics Devices

127

Let us look at these three requirements one at a time. Metadata is information about a data file that is stored with the data file and can be accessed by external programs for various uses. Metadata can be entered manually, such as with editing of content where a description of a picture or a scene is added to describe it. This is the most common way to generate metadata today (aside from date stamps and other simple automatic metadata), and because of the effort required, the creation of metadata today is rather primitive. Better for most users is automatic metadata generation, where the information about the data file is derived from the content in the file. Such technologies are being developed, and they will revolutionize the usefulness of new and historical content. With such a powerful metadata creation tool, a user will be able to create metadata on all of this historical content as well as new content, greatly increasing the information available for rapid access of content. For text, metadata could consist of key words or phrases that give the gist of the material. For audio, metadata could be key words or phrases derived from a converted version of audio as well as useful descriptions of nonhuman sounds. For still or moving images, metadata could include image elements that allow recognition of people, animals, or places and for the moving image some sort of synopsis of what action occurs involving which elements. For all these data files, date and location information (if available from, e.g., GPS) could also be useful information. Indexing of content uses metadata and externally generated key words about individual files to speed up a search of this content. If we can extend this capability to include being able to abstract the context of different pieces of content, then these tools could become even more useful. Much investment and research is going into developing such capabilities. The benefit to users will be enormous. The Future Home Media Center As a centralized repository and access point for all of a home’s user-generated and commercial content, a future home media center must •

Be easy to use.

•

Be able to provide all required storage and sharing services.

•

Have the bandwidth and storage capacity to meet household needs.

•

Have easy to install additions of storage and bandwidth.

www.newnespress.com

128

Chapter 5

•

Have software capable of providing proper control of content access, security, and privacy of content.

•

Provide automated protection of the content for backup restoration from the range of simple file corruption to complete home disasters where the home media center and its local backup itself may be destroyed.

•

Be inexpensive enough in volume that it can achieve a mass market.

Even if the general architecture that is called a home media center does not become the standard way that content is handled in the home, these requirements still hold. In order to incorporate these capabilities, various technologies will likely be integrated into the centralized home storage repository and control center: •

DRM or other content security as appropriate.

•

Means to protect personal content privacy such as storage device encryption (proposed by the Trusted Computing Group).

•

Networking capability providing adequate bandwidth to support many streams of ever higher resolution content in the home (including wireless networking).

•

Storage devices providing the proper storage capacity at the point of use and access to the centralized content so this data can be shared, refreshed, and controlled.

•

Automatic metadata creation of text, audio, and video content, including context placement capability to enable rapid search.

•

Online services offering backup, protection, sharing, and rapid recovery of home content.

If these features can be provided in a box that could cost less than $300, then in volume the market would be very large. This may be possible with some additional years of electronic integration and application maturity. Another option is to include such a product into a monthly service fee, including the home hardware as well as the online backup service. This is likely to be an approach taken by service providers such as telcos, satellite providers, or cable companies.

5.3.5 Backing Up and Disaster Recovery for Home Media Centers Home media centers or other home content devices will contain some of the most valuable information in a home. A centralized repository for family digital photographs,

www.newnespress.com

Storage in Home Consumer Electronics Devices

129

videos, and other personal content as well as DVR recordings and purchased and downloaded content from the internet and other sources, has a great value to home users. The loss of some or all of this information can be a nuisance at best, and in the case of user-generated content the irrecoverable loss of content would be tragic—like losing a family photograph album in a fire! For this reason, it makes sense that a home media center as a likely aggregator for valuable personal and commercial content should have a means of backing up the data it contains as well as protecting this content from a disaster that may destroy the digital storage devices where this content is stored in the home. Backup of a home media center should be an automatic function of the device, and once set up it should proceed without frequent intervention by the user. The setup itself should be easy to do, so that even a nontechnical user can quickly implement the most common basic data protection with capability for more experienced users to do more if they wish (although this has its own dangers, since if nontechnical users are forced to use this function and it is not well designed, they could get lost and even do damage). Ideally, as soon as a device is connected to the network where a home media center exists, it should back up that data and be able to restore it to the original device (or its replacement) when needed. The home media center should also make it easy to move content from one device to another, such as when hardware is upgraded or changed. As should always be the case, this content should be secure from outside intruders through encryption and access control. A further step in the protection of user content (both personal and commercial) is to create an online service where the content can be automatically backed up to a secure location on the internet. This location must be secure and prevent unauthorized access to the content. It must be set up to do incremental backups of the content on a regular basis and, if needed, individual file, or even entire household data recovery, should be easy to perform.

5.4 Chapter Summary •

Digital video recorders (DVRs) have become a common home consumer device. DVR expansion and content sharing have led to efforts to create external direct attached and network attached storage systems to provide additional storage capacity to these devices as well as allowing content sharing. eSATA is very

www.newnespress.com

130

Chapter 5 attractive for storage expansion of DVRs since the current standard supports up to 3 Gbps data rates and daisy chaining of eSATA storage devices, allowing additional storage expansion.

•

Many different options for home media networking to support sharing of DVR content are likely, including coaxial cable-based systems, wired and wireless Ethernet systems (e.g., 802.11x wireless networking), and possibly power or phone line-based networking systems. The choice will depend upon the devices being connected, what is supported by one’s service providers, and local connectivity options.

•

A key reason for a home network is backup and protection of data and content in the home. In addition to in-house backup, internet-based service providers are developing utilities that can back up or synchronize the data on computers and other devices in the home. Such remote backups can provide a path to disaster recovery of personal, business, and commercial content in case of an accident that damages or destroys the devices on which the original content was stored.

www.newnespress.com

CHAPTE R 6

Storage in Mobile Consumer Electronics Devices Objectives in This Chapter •

Look at the functions, design trade-offs, and uses of digital storage in common mobile consumer electronics products.

•

Examine the unique and demanding environment of digital storage for entertainment and navigation in automobiles, and how to design systems made for these environments.

•

Explore the trade-offs in performance and power for mobile music and video players, and how new developments and capabilities could lead to richer content available for these devices in the future.

•

Learn what sort of storage capacities will be required for mobile players in the future, and how to choose the best type of storage device from the mobile storage hierarchy.

•

Find out how digital storage is integrated into still and video digital cameras, AV players, mobile phones, and other mobile consumer products.

•

Consider what may be required for a personal life-log device that records your life experiences as they happen.

6.1 Introduction This chapter extends the analysis of the prior chapter into the design and digital storage requirements of consumer devices in the home to mobile consumer devices. Applications such as automobiles, mobile music and video players, and digital still and

www.newnespress.com

132

Chapter 6

video cameras have significant challenges in environmental reliability, battery life in normal use, and resolution requirements. These characteristics must be taken into account to design the right type of digital storage for these applications. This chapter explores the unique drivers in mobile consumer product digital storage, both at the present time and for likely developments in the future. It gives some insights into what sort of storage capacity will be available and how much this will cost for the available storage options. It ends by briefly considering new uses of digital storage in mobile devices that could increase the required storage considerably both in these devices and in the home.

6.2 Automobile Consumer Electronics Storage Automobiles represent one of the most challenging environments in consumer electronics. I was raised in the upper Midwestern United States and I experienced temperatures that ranged from as low as −30ºF (−34ºC), not including wind chill, to as high as 125ºF (46ºC). In a car, temperatures can get even higher with the windows rolled up. Furthermore, electronic products designed for this environment must also deal with the regular vibrations, accelerations, and shocks that are common to automobiles. Storage devices designed for these applications must be very robust to survive in such an environment. Because of the long life expected for embedded automobile components in such a harsh environment, qualification times tend to be rather long, often better measured in years rather than months. This section will explore the requirements for automobile applications, as well as storage devices designed to be used in automobiles. Common storage devices used in automotive consumer applications are optical discs, hard disk drives, and flash memory.

6.2.1 Digital Storage for the Automobile Table 6.1 shows some characteristics that could be expected for storage devices in an automobile environment. These are not comprehensive, and many of these environmental conditions such as shock and vibration are reduced at the location of the electronic device to increase the reliability. Automobiles are incorporating more and more electronics. With the rising price of gasoline, it is likely that within the next 10 years most automobiles will be run at least

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

133

Table 6.1: Automobile Environmental Conditions of Interest for Most Consumer Electronics Applications Environmental Factor

Range

Temperature Relative Humidity Vibration Shock

−40°F to 120°F (Internal Automobile Temperature) 10% to 100% Can Be Very Severe, Usually Damped Around Electronics Can Be Very Severe, Usually Absorbed Around Electronics

partially off of electricity rather than combustion engines. Electronics is making its way into automobile control systems, safety and diagnostic systems, as well as navigation and entertainment systems. Most digital storage in automobile control, diagnostic, and safety systems is embedded solid state memory since these storage requirements are generally not too great and the robustness required is high. Automobile navigation and entertainment systems are the exception, because they could require very large amounts of storage. This storage could be removable or fixed, depending upon the purpose it is used for. Navigation systems use digital storage devices for two basic purposes: 1.

To supply data to the navigation system of basic routes, locations, and maps.

2.

To acquire and store real-time traffic updates and suggest alternative routes when appropriate.

The first approach uses relatively fixed data that changes slowly with time. Data for these applications is often supplied on removable media such as an optical disc (a CD or DVD). Alternately, it could be downloaded into a mass storage device in the automobile that may use a hard disk drive or a flash memory. Depending upon the resolution of the maps, route, and destination information, an entire country could be represented in basic maps of different locations on one or two DVD discs (or one blue laser DVD). If greater resolution maps are required, then the storage requirements can increase even further. If audio or video clips about possible destinations are included, then the storage capacity for even this relatively static content can become rather large. It is easy to see that a simple static navigation system could grow to be many gigabytes in size. As the storage capacity of optical storage (or a mass storage download location) grows, such navigation could become much richer, becoming to some extent part of the

www.newnespress.com

134

Chapter 6 Table 6.2: Comparison of Digital Storage Requirements for Various Resolution Maps

Description of Map California Highway Map (PDF) Santa Cruz County, California Topographic Base Map Salinas, California Area Landsat Image

Storage Capacity Needed (MB)

Source

7.73

http://www.metrotown.info/

9.8

http://www.usgs.gov/

∼55

http://www.nasa.gov/

entertainment system. Table 6.2 gives some idea of the amount of storage required to store various resolution maps of a geographic region. Real-time traffic update systems for automobile navigation have become very popular in Japan due to the narrowness of the streets in Tokyo and the propensity for traffic to become congested. For this reason, Vehicle Information and Communications Systems (VICSs) have become very popular. Several million of these systems have been installed in automobiles in Japan and other Asian countries. These in-car navigation systems are combined with a sensor and radio network located on the streets of the town that give real-time traffic condition information and with the help of the navigation system provide the user with alternative routes to avoid congestion and delays. The real-time route updates are stored on an onboard storage device. Usually these storage devices are ruggedized hard disk drives that also contain map, route, and perhaps destination information. These hard drives allow for rapid updating of the route information as well as ongoing changes in the general map and destination information. It is likely that hard disk drives could be augmented by a flash memory write cache (a hybrid drive) or even substituted by a solid state drive in future navigation systems depending upon the storage requirements and price point of the application.

6.2.2 Basic Layout of Automobile Navigation and Entertainment System Let’s look at some examples of schematics for an automobile entertainment and navigation system. Figure 6.1 shows an Floating Point Gate Array (FPGA) implementation of a multimedia backseat entertainment system. This system is capable of multiple content displays simultaneously. The Composite Video Blank and Sync (CVBS) analog signal is converted to an ITU656 digital signal by a composite

www.newnespress.com

Touch Panel Control

12C Control

Keyboard

Platform IP Cores User IP Cores

GPIO

FPGA ASSPs

Display logiCVC Video Control

Display

Video Camera VCR

Connectors

logiCAN logiUART CAN/Lin

Ulti/WIN Versatile Video Input

CVBS Decoder

OPB CoreConnect FPGA IP Interconnection BUS

User IP (MPEG Decoder)

Ulti/WIN Versatile Video Input

User IP (MOST Control)

logiBITBLK Graptic Accelerator

UltiMEM UMA Memory Controller

SDRAM DDR Flash

Figure 6.1: Backseat Car FPGA-Based Entertainment System Schematic.1 1

After Xilinx block diagram.

SDRAM (DDR)

Flash

135

www.newnespress.com

DVD

MicroBlaze PPC, ext CPU

Storage in Mobile Consumer Electronics Devices

Drivers

logiCVC Video Control

136

Chapter 6

video signal decoder, an ASSP component. The MOST MAC (Media Access Control) and MPEG decoders are customer-added IP cores. The I2C (Inter-IC bus) and GPIO (General Purpose Input Output) IP cores are used for the keyboard and touch controller interface. The UltiMEM IP core is configured for unified memory architecture and six access ports. The 32-bit DDRAM (Double Data Random Access Memory) ensures 800 MBps bandwidth for display refresh, video streaming, graphics acceleration, and CPU program execution. The logiCAN and logiUART-local interconnect network (LIN) IP cores are used for network connections with in-car body electronics. A GPS navigation system reference design schematic is shown in Figure 6.2. This system uses CD/DVDs for the map information, but other storage devices could be substituted for the optical media. SAW Filter

Antenna

SAW Filter

LNA

ADC

IrDA Clock Driver

VCO

USB MMC SmartCard

Frequency Synthesizer

OSC

DSP/CPU LCD Controller

Keypad

EEPROM Flash SRAM

Touch Screen Controller Audio Codec

Speaker Headphones Microphones

Main Supply LDO

RF Supply LED Supply LCD Supply LDO

DC/DC Boost Converter

DC/DC Boost Converter

Digital I/O DSP/CPU Core Analog Supply Memory Supply DC/DC Buck Converter

LDO

LDO

Supply Voltage Supervisor

Power Management

Integrated Power Management

LEGEND Processor Interface Amplifier Logic

Battery Power Management

Power ADC/DAC Other

Figure 6.2: GPS Receiver Block Diagram.2 2

Courtesy of Texas Instruments.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

137

The Global Positioning System (GPS) works on the principle that if you know your distance from several locations, then you can calculate your location. The known locations are the 24 satellites located in six orbital planes at an altitude of 20,200 km. These satellites circle the Earth every 12 hours and broadcast a data stream at the primary frequency L1 of 1.575 GHz. This data stream carries the Coarse Acquisition (C/A) encoded signal to the ground. The GPS receiver measures the time of arrival of the C/A code to a fraction of a millisecond, and thus determines the distance to the satellite. The major components of the GPS receiver are: •

Front End: The GPS L1 signals (Maximum = 24 signals) at 1.575 GHz are received at the antenna and amplified by the Low-Noise-Amplifier (LNA). The RF front end then filters, mixes, and amplifies Analog Gain Control (AGC) to bring the signal down to the IF frequency where it is digitally sampled by an Analog to Digital Converter (ADC).

•

Digital Signal Processor/CPU: The ADC samples of GPS C/A code signals are correlated by the Digital Signal Processor (DSP) and then formulated to make range measurements to the GPS satellites. The DSP is interfaced with a generalpurpose CPU that handles tracking channels and controls user interfaces.

•

Memory: The processor runs applications stored in memory. The operating system is stored in nonvolatile memory such as EEPROM/flash/ROM. Applications may be loaded in flash or DRAM.

•

User Interface: The user interface allows the user to input/output data from the receiver using input commands via microphone and touch screen, and output MP3 to the earplug.

•

Connectivity: Connectivity allows the receiver to connect to the USB port.

•

Power Conversion: Converts input power (battery or wall plug) to run various functional blocks.

6.2.3 Storage Device Trade-offs and Options for the Automobile Automobile entertainment systems typically use optical disc formats such as CDs and DVDs3 today to store music and movies, but this could change in the future if 3

The use of consumer DVDs in automobiles required the addition of the Automotive DVD Playback System to the DVD Copy Control Association’s Content Scramble System.

www.newnespress.com

138

Chapter 6

consumers demand large amounts of onboard mass storage to store movie libraries and other information. Also, as mentioned earlier, destination audio and video information could be available through a navigation system to learn more about a destination before reaching it. A real-time navigation system requires digital storage media that can be written and read from many times. These are popular in Japan where sensors in the streets detect local traffic and road conditions and these signals are used to tell subscribing automobiles what route should best be taken to reach a destination. Real-time navigation systems, as well as systems that require lots of content for entertainment or navigation, often use a storage device such as a hard disk drive. Hard disk drives used in automobiles live in a very harsh environment, very different from those in computers and even most consumer applications. In order to operate under the temperature extremes, vibration, and shock conditions found in automobiles, the track density and linear density of hard disk drives are reduced to give lower storage capacity but a more rugged storage product with performance margin to spare. For the very extreme temperatures in which hard disk drives are expected to operate, they must be heated or cooled to stay within a specified environmental tolerance. A shock absorbing case can be used around a hard disk drive to decrease shock and vibration sensitivity further. There have been several designs of such cases. One that has been promoted by several Japanese companies is the Information Versatile Disk for Removable Usage (iVDR) standard.4 Figure 6.3 shows a product using this design. The iVDR provides a modest amount of additional shock protection beyond what the initial disk drive would have, and allows for a removable storage option using a hard disk drive. Although the most popular storage media for automobile applications is currently optical discs, hard disk drives and flash memory are starting to show up in automobiles as well. Appendix A compares important environmental and performance specification differences for several 1.8-inch hard disk drives as well as a flash memory solid state drive. Although the flash memory device is more robust at extremes of temperature and shock, the price per gigabyte is three to four times higher, and likely to remain so for several years to come. Although the various power specifications for the hard disk drives appear to be higher, in actual use, the hard drive stays on less than five percent

4

See http://www.ivdr.org.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

139

Figure 6.3: iVDR Storage Device Cartridge and Reader. of the time (low duty cycle), filling a semiconductor buffer memory that is used for actual playback in an entertainment system. As a consequence, the total power usage is not very different for many flash based devices than for a hard disk drive based equivalent device.

6.2.4 Road Map for Automobile Digital Storage Requirements With the increasing price of gasoline, it is likely that electronics will become more important in automobile design. By the next decade, it is likely that there will be a substantial population of automobiles that may use electric power sources for all or part of their motive power. With electronic controls and power sources increasing in their capability and numbers, the integration of other functions into the electronics for navigation, entertainment, accident and liability records (automotive black box), and so forth will be natural additional capabilities. Over the next couple of decades it is even likely that automated driving (automotive autopilot) could move to the consumer sector. There are many successful university autonomous driving programs that may find implementation in military and dangerous environments in the next few years. Once such autonomous driving systems go into consumer transportation, riders will have time to do more than just watch the road while they drive. This opens up even more need for communication and entertainment technology to use this time productively or for entertainment. These changes and the

www.newnespress.com

140

Chapter 6

increasing amount of time that many people spend in their cars will drive new requirements for digital storage. Real-time traffic and road updates using a wireless network or radio system and a rewritable storage device such as a hard disk drive will become more widespread, particularly for autonomous vehicles or autopilots. As an aid in navigation and as a source of information to the passengers, it is likely that information on locations, commercial information in the area, and other information automatically recorded in an automobile as it approaches a destination, will become more common. Base maps of roads will probably consume considerably less than 1 GB and cover a very large area and a fair amount of detail. As indicated in Table 6.2, topographical information and satellite images would require much larger amounts of digital storage. Satellite and other high resolution images and even video taken from a location may also be uploaded on automobiles. Combining these needs with traditional entertainment, especially HD video, could require fairly significant storage either in the automobile itself or available to it through wireless networking. It is not difficult to see a need for several hundred gigabytes or even a terabyte in an automobile navigation and entertainment system in the next few years. When automobiles are parked near to their base (your home), they will need to connect to your home network, probably using a wireless network interface. This will allow updating and synchronization of content between your home and your automobile. The automobile will likely only contain copies of a subset of your entire data, since the larger bulk of your content is available relatively conveniently. If DRM issues can be resolved, it would be very convenient for consumers to be able to copy their CDs and DVDs to their home network to make them available in their automobiles without having to handle physical disks that can become damaged and lost. A much larger library of content can be used and organized for easier use when they can be put on a network storage device. Thus in the long run I expect that optical media will not be the preferred way to use content in automobiles. Hard disk drives and flash memory are more likely to be the primary storage in automobiles in the future. Flash is attractive because it has a greater operating temperature range and it is more shock and vibration insensitive. However, it appears likely that flash memory will remain about two to four times more expensive than an equivalent storage capacity available on a small form factor hard disk drive. For large content storage, even in an automobile environment, a role will probably remain for hard disk drives.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

141

Flash memory may be a way to transfer content in and out of a car (like a USB drive) or be used for certain applications within the automobile that need a more robust storage product. I also consider it likely that hybrid storage devices that contain a flash memory cache built into a hard drive or on the car electronics motherboard could provide some of the faster response under extreme environments that flash can provide, while adding only a little to the cost of the hard disk drive.

6.3 Mobile Media Players MP3 and video portable players are some of the most popular consumer devices. These products store local copies of music, photographs, or video content and so digital storage is an important element in their design. In this section we will describe the trade-offs in design and performance that must be made in battery powered consumer devices. Many of the observations made here will also be applicable to other mobile devices.

6.3.1 Mobile Media Player Designs We will investigate two types of portable media player designs and the digital storage and other design requirements for these devices. These two player devices are (1) mobile music players and (2) the more powerful multimedia player devices that also play video content. Both of these devices can store personal as well as commercial content. Basic Layout of Mobile Media Players A media player is a simple device in concept. It consists of a sound system that is often an amplifier with a headphone jack, connected to a codec that decodes data read back from a storage device. Typically, a battery provides the power for the device. Since the power from a battery is limited, power consumption is a critical constraint in the amount of time that a customer can use a device between charges. Thus power saving modes and various other techniques are used to reduce power consumption and increase the time that the device can be used on a single charge. In addition to the power requirements, customers often prefer to have devices that are more compact, have a lighter weight, are more durable under likely usage conditions, have an easy way for

www.newnespress.com

142

Chapter 6

users to obtain content in the device, look good, and of course are within the budget of the user. Most mobile portable media players are temporary storage devices that contain a copy of the content that is copied from some other device owned by the consumer, such as a computer. Since the player can be lost or damaged, it is wise to retain the original copy on a static device that stays in the home of the user, where it is hopefully less subject to damage and loss. Figure 6.4 shows a block diagram of a mobile music player (with voice recorder capability) and Figure 6.5 is an image of a teardown of a portable music player. The MP3/Voice Recorder-Player performs noise reduction, speech compression, and MP3 stereo recording/playback at all rates and options required by the MPEG/speech compression standards. Following is a description of the components for such a device.

LCD Display

LCD CTLR

LED

ROM GPIO Keypad

Keypad

EMIFS

Flash Storage

Stereo Headphones DSP Mono Speaker Microphone Input Stereo Audio/ Audio Line_in

Battery

125

Audio Codec

SPI

USB

USB

McBSP

McBSP

Main/Core

Audio CODEC/ Audio Amp

LED/ LCD Supply

I/O Supply

DC/DC Step-Down Converter

Linear Regulator (LDO)

DC/DC Boost Converter

Sync Boost Converter

MMC/SD

Smart Card MMC SD

LEGEND

Charger

Power Management

Figure 6.4: MP3 Player and Voice Recorder Block Diagram.5

5

Courtesy of Texas Instruments.

www.newnespress.com

Processor Interface Amplifier Logic Power ADC/DAC Other

Storage in Mobile Consumer Electronics Devices

143

Figure 6.5: 8 GB iPod Nano Teardown Showing Flash Memory Storage.6 •

The DSP performs the audio encode functions, executes post-processing algorithms like equalization and bass management, and system related tasks like file management and the user interface control.

•

The memory stores executing code plus data and parameters such as content metadata.

•

The peripheral interface allows the user to control input/output (I/Os) and the display.

•

The audio encoder/decoder or codec interfaces with microphone, radio signals (if the device included a radio receiver) and other audio input with the headphone and possible speaker for audio output.

Power conversion changes the provided power to run various functions in the device. Basic Layout of a Personal Multimedia Player Music players have been the most popular of the mobile players so far, but with larger storage capacities available at low cost video players are becoming increasingly powerful and affordable. Multimedia players that play commercial video as well as audio content and can also be used to show user generated content such as digital pictures or home videos are becoming more popular and will probably eventually

6

Image courtesy of iFixit.

www.newnespress.com

144

Chapter 6

displace many of the music only players. Figure 6.6 shows a block diagram of a portable media player that handles video content. Figure 6.7 shows images of a teardown of a personal multimedia player. The Personal Media Player (PMP) is a handheld audio/video system that can record and play back Audio/Video (A/V) from TV, DVD player, camera, or a media file downloaded from internet. A description of the various components in this block diagram is given below:

L

Stereo Audio Line In

Stereo Audio Line Out

R Touch Screen Controller

R

LCD Interface

L FM Radio IC

L

Stereo Audio Codec

R

Serial Interface (12C, SPI, 12S)

LED Backlight

Stereo Headphones

Touch Screen Controller

Speaker Microphone In

LCD Panel

Keypad

IR Remote Control

RTC

CCIR 656-Bus Video Decoder

Video In

MPU Video Buffer

Lens CCD/ CMOS Sensors

Digital Camera Front End Signal Camera

Buffer

DDR2

NTSC/PAL Digital Media Processor

RS232 RS485

Clock

USB Hard Disk Drive

Battery Power Management

SD/MMC/ MS Interface

Compact Flash Adapter

Flash Memory

JTAG

Primary Supply

LED Backlight Supply

LCD Supply

System Power

DC/DC Boost Converter

DC/DC Boost Converter

DSP/CPU Core Supply DC/DC Buck Converter

LEGEND Processor

I/O Supply

LDO

Supply Voltage Supervisor

Integrated Power Management

Interface Amplifier Logic Power ADC/DAC

Power Management

Figure 6.6: Portable Personal Media Player Block Diagram.7

7

Courtesy of Texas Instruments.

www.newnespress.com

Other

Storage in Mobile Consumer Electronics Devices

145

Figure 6.7: Teardown Images of Sony Walkman (Hard Disk Drive Based) Personal Media Player Identifying the Various Components.8

8

•

The DSPs and CPU (the digital media processor) performs digital audio and video processing for all standard media formats. The CPU executes operating system software and controls data transfers and user interface.

•

The digital video input comes from the analog video decoder or the CMOS/CCD camera. The analog video decoder converts the NTSC/PAL/S-Video analog input to raw digital video.

Courtesy of Portelligent, Inc.—teardown.com.

www.newnespress.com

146

Chapter 6

•

The stereo audio codec converts the audio input from microphones, FM receiver, and stereo audio line sources into digital audio.

•

The CCIR656 Bus moves raw digital video content to the DSP.

•

The serial interface enables file transfers with the host computer via standard interface formats such as USB, RS232, and JTAG interfaces (this could eventually include serial ATA).

•

The memory interface handles direct memory interfaces with various portable memory media including compact flash cards, SD cards, MMC cards, memory sticks, and so on.

•

The IDE interface handles data transfers with a hard disk drive.

•

The user interfaces allow users to control the personal media player using a keypad, IR remote control, voice, or touch screen.

•

The power management and conversion system converts input power from a battery or wall plug to the various functional blocks of the PMP.

6.3.2 Display Technologies in Mobile Devices Creating inexpensive and easy to use personal high resolution display technologies has implications in power use, physio-psychological effects, as well as weight, convenience, and price. Nevertheless, we expect that there will be considerable effort expended to make higher resolution content available to mobile users. Figure 6.8 shows a display technology road map (including mobile or wearable devices) through 2015 (after a table in an Optoelectronic Industry and Technology Development Association Report). Based on these road maps we might expect initial products with a useful mobile high resolution display device to be available sometime toward the end of the decade and this will drive requirements for higher resolution digital content and thus higher capacity storage devices. A very interesting option for getting a large format display from a mobile battery powered device is the pocket projector technology from Novalux.9 This technology uses surface emitting infrared lasers and frequency doubling to achieve 1.5 W scanning RGB 9

See http://www.novalux.com.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices Year 1998

2000

2005

Wearable

2010

2015

Wider Angle, Higher Resolutioin

HMD FMD

147

High Reality

Size and Weight Reduction

Wristwatch Display (Wrist Visual Phone) Less Power Consumption LCD Scanning

3D Goggle p-Si TFT LCD Si TFT LCD Reflection Type LCD CGS LCD Ferroelectric LCD

LCD Video

Reflection Type LCD

Flat TV

Note-PC

Flat PC monitor

CRT CRT

Film Display Devices (LCD, EL)

Organic EL Hybrid Film LCD

Ferroelectric LCD

CRT

Digital Integration TV

PDP LCD

Paperlike Display

Sheet PC

Si TFT LCD

a-Si TFT LCD

Home, Office and Personal

Intelligent Element

Organic EL

Card TV

p-Si TFT LCD

Film LCD

••

Mobile Tool Mobile Network PC E-Book E-Newspaper, E-Magazine

PDA Cellular Phone

Mobile

LD Scanning

Multifunctional Display Goggle-less 3D

LCD, PDP, PALC

PDP, PALC, LCD PDP, PALC, Si-LCD

Multifunctional 10

Public Indoor

Home Theater Teleconference Classroom, Conference Hall E-bulletin board Electronic Advertisement, Illumination

Outdoor

Stadium CRT CRT

Wall Display

LCD Projector

Exhibition Hall

High Reality

Video Wall 3D

ILA, DLP Projector LED

Figure 6.8: Road Map of Electronic Display Technologies.10 lasers that can provide 100–300 lumen projection of images from battery powered devices. This display technology automatically projects and focuses on any surface allowing very interesting potential applications. Figure 6.9 illustrates how such a device can be used. The company says that this technology should be introduced in cell phones, PDAs, and other mobile devices by 2009. From a storage point of view, the pocket projector is interesting since it allows watching larger format content on a mobile device than is possible using the small LCD screens typical on PMPS, PDAs, or cell phones. This translates into a capability of watching higher resolution content from a mobile device. Higher resolution content translates into larger files and that means larger capacity storage needs. The storage required could be embedded in the product or else available on a local network, perhaps in personal network storage architecture such as Seagate’s DAVE or LSI’s Blue Onyx. 10

After Optoelectronic Industry and Technology Development Association Chart.

www.newnespress.com

148

Chapter 6

Figure 6.9: Demonstration of Uses of Novalux Pocket Projection Technology.11

6.3.3 Power Requirements for Mobile Devices Several factors need to be taken into account in comparing the power used for a flash based and a hard disk based media player. •

The required data rate for continuous streamed content.

•

The data rate out of the storage device.

•

The power available from the battery of mobile power source.

•

The power budget for the mobile device including storage, electronics, audio player, and display technologies.

In mobile devices power is a serious design parameter. All the components in a mobile player require power. The power requirements for different system components (after 11

Image courtesy of Novalux Corporation.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

149

evaluating and optimizing performance and usage life trade-offs for the expected products applications) can be used to create a power budget that then defines the mobile power supply needs of the device. Figure 6.10 gives us some insight into the process of evaluating the actual power usage of a personal multimedia player. The data rate of a storage device such as a hard disk

Hard Disk Drive

Application Speed

Native Speed

MP3 or MPEG

Buffer FAST

SLOW

Select the Minimum Size Buffer to Meet the Power Goal and Minimize Cost.

MP3 Player Hard Disk Drive Access Example and Current Consumption One Sector Read

Peak Current

Active Phase: 20 Seconds Power On

Idle Phase

Power Off

Power Cycle: 20 Minutes

Example: Disk Drive 300 mA @ 3.3 V, 3% duty cycle, 9 mA average

RAM Buffer 15 mA @ 3.3 V

MP3 Player 70 mA @ 3.3 V

Total Power 310 mW

Figure 6.10: A Buffer Can Be Used to Match the Content Delivery Speed of a Storage Device Such as a Hard Disk Drive and the Application.12

12

After Duncan Furness, Agere, “Architecting Minature Hard Disk Drives for Battery Savings,” Planet Analog, 2005.

www.newnespress.com

150

Chapter 6

drive (or a flash memory device) is much greater than required by an MP3 or MPEG player application. A buffer is used as a host speed matching device. The size of the host buffer will determine the power dissipation; more buffer means less power-hungry spin up/down drive cycles, resulting in lower average power dissipation. Trade-offs in buffer size, power, and cost help host manufacturers select a buffer that meets the power dissipation goals of the application while minimizing cost. Typical music player buffers are in the range of 8 to 16 MB. For instance, an MPEG-2 video file streams at less than 3 MBps while a 1.8-inch hard disk drive has a sustained data rate of 16 MBps (128 MBps). Since the hard disk drive consumes a fair amount of power while reading (1.1 W) we would probably design a battery based hard disk drive media player so that the disk drive would not remain on all the time. Instead, such a media player would turn the disk drive on only a fraction of the total player power-on time (less than three percent). As a result, as shown in Figure 6.10, the total power used in this player would probably be close to 300 mW rather than the 1.1 W if the hard disk drive were constantly running (about a 70 percent power reduction). Note that flash memory designed devices may also use RAM buffers for play out since their data rate can exceed that required for the streaming content. By the way, it is interesting to note that almost half of the total power consumed by a hard disk drive is used in the electronics rather than the electromechanical components. Power is one of the major constraints on mobile consumer electronic products. Conventional battery technology lasts no more than a few hours and promising technologies such as fuel cells have not yet shrunk to the point where they can be used in many mobile devices and consumer electronics companies and consumers have yet to be convinced that fuel cell powered devices will be safe or reliable. Battery technology has been developing over time as shown in Figure 6.11 with battery power density roughly doubling every 10 years. Fuel cells will likely find some application in mobile devices by 2009–2010, perhaps as mobile rechargers if not the actual mobile power source. Many of these devices use methanol as the source of hydrogen, but there are some technologies that directly use hydrogen. Some issues needing resolution for the use of fuel cells include heat management and exhaust of by-products (such as water). Fuel cells could increase the energy density of mobile energy sources considerably. Another interesting possibility is wireless power to recharge mobile devices. At the 2007 CES Conference, Herman Miller (a designer of office furniture) showed the

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

151

Volumetric Energy Density Wh/l

700 600

Doubling energy density in ten years!

580 525 490

500 450 400 300 200 120 100 ’85

Li-ion 390 330360 350 315 330 280 300 235 240 Ni/MH 210 200 230 180 190 180200 160 ’90

’95

Year

’00

’05

Figure 6.11: Battery Energy Volumetric Density Development.

eCoupled technology developed by Fulton Innovations.13 This technology has a resonant circuit connected to the power supply for the mobile device that can resonate with a similar circuit in the furniture with the capability of transferring small amounts of power to recharge the battery of mobile devices. The technology only operates at a distance of about an inch, so the mobile device would probably have to sit in some sort of cradle. If such technologies became common in commercial furniture, it could help extend the life of mobile device batteries and help encourage more feature rich applications. MIT demonstrated at the 2007 CES their “Witricity” initiative which has been shown capable of transferring 60 W of electrical power 7 feet at 40 percent efficiency. With more power available for mobile devices, features such as higher definition content projection (i.e., with Novalus personal projectors) may be possible, allowing viewing of such content and hence a requirement for greater storage capacity and read bandwidth to support playing such content.

13

See http://www.engadget.com/2007/01/19/herman-miller-planning-desk-of-the-future/.

www.newnespress.com

152

Chapter 6

6.3.4 How to Choose the Right Digital Storage for a Mobile Media Player In addition to power there are other characteristics that are important to the use of one digital storage device versus another. These include: •

Price for a useful capacity (always very useful).

•

Size and shape.

•

Weight.

•

Ruggedness (shock, vibration, and other environmental sensitivities).

•

Heat generation.

We shall look at how choices can be made for each of these characteristics.

Price ($)

Figure 6.12 compares the price and storage capacity of flash memory and various form factor hard disk drives as of mid-2007. Flash memory devices have a minimum base price of $7 to $10 while the hard disk drive minimum base price is in the $40 to $50 range. Flash memory can be extended to larger storage capacities by adding more NAND flash chips with an increase in total price. Flash capacity increments are smaller than hard disk drive increments since each disk surface adds so much more capacity. We can see that for lower storage capacity requirements flash memory will be cheaper, $100 $90 $80 $70 $60 $50 $40 $30 $20 $10 $0 0.001

2.5-inch 1.8-inch 1-inch Flash

0.1

10

1,000

Storage Capacity (GB)

Figure 6.12: Comparison of Price and Storage Capacity of Flash Memory and Hard Disk Drives in Mid-2007.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

153

while if the storage capacity requirements are large, hard disk drives can be cheaper. Of all the other factors, the price to buy needed sufficient storage for an application is paramount for most consumers. The figure also shows that there is a range of storage capacities that hard disk drives cannot address due to the incremental nature of hard disk drive capacity increases. Hard disk drives have a definite length and width dimension defined by the size of the disk form factor used. The height is a function of the dimensions needed for the magnetic head and drive electronics. Flash memory is composed of collections of semiconductor chips and so the options for packing in storage capacity are greater and devices using flash memory can be more flexible in their dimensions. This can provide definite advantages for many applications if all other factors are equal. The weight of flash memory is dependent upon the number of flash memory chips required as well as the packaging. Higher flash memory capacities will weigh more, but the weight generally is less than for a hard disk drive device unless the flash memory is very large. Likewise, flash memory is generally more rugged than hard disk drives in terms of operating shock. However, as indicated earlier, the drive may be in a low power nonoperating state most of the time (more than 95 percent of the time) where the device plays out of a RAM buffer. The nonoperating shock specifications for small form factor hard disk drives are about those of a NAND flash memory device. If an accelerometer is included in the hard disk drive or the application device then even the operating shock specification can be improved further, since the drive could detect when it is in a free-fall state and retract the heads to their protective nonoperating state. Vibration and temperature sensitivity can be lower for flash memory than for hard disk drives. Flash memory may also work better in many moist or even wet environments than hard disk drives. In general, since the power usage for a hard disk drive consumer device can be similar to that of a flash memory device (with the proper power management features) the heat generated may be about the same and the battery life may also be similar. In addition to these characteristics other criteria may be important for various consumer applications in the choice of a storage device. For instance, flash memory cells will wear out with multiple writes and the intrinsic read speed of flash memory is less than the write speed (this is also true of hard disk drives but with flash the difference is greater). These characteristics of flash memory may cause a designer to not use flash

www.newnespress.com

154

Chapter 6

memory for write intensive applications such as DVR functions or real-time traffic navigation. Let us look at the particular case of why Apple switched from using 1-inch hard disk drives to NAND flash memory in their music iPod product line. At the time the decision was made, 1-inch hard disk drives had larger storage capacity for the price compared to flash memory, but the storage capacity in a flash memory device that could be purchased for an affordable price by many consumers (less than $300) was approaching a point that many MP3 songs could be stored on it. Since a flash memory player would be lighter weight and could be made thinner as well as smaller in other dimensions than a hard disk drive based device, the “portability” of the device would be better. Note that in general it is advisable to make a mobile consumer device with a specific gravity close to that of water (which is also close to our specific gravity) so they would not seem to weigh too much when on our bodies. As shown in Figure 6.13 the flash based iPod Nano was more “portable” than the iPod

? Songs

Portability

? Size

Shuffle Nano Mini iPod

Songs

Figure 6.13: iPod Portability Versus Song Capacity.14 14

Courtesy of Semico Research Corp.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

155

Mini (with a 1-inch hard disk drive) even though it couldn’t carry as many songs. In addition, with the declining price of flash memory, a player with 10,000 or more tracks would be possible with flash memory in an attractive price range within a few years from the introduction of the iPod Nano in 2005. The combination of greater portability and adequate storage capacities for the application probably played a significant role in the decision to switch the Apple music players to flash memory. It should also be mentioned that Apple was able to make contracts with Samsung and other large flash manufacturers to get its flash at considerably lower prices than they would be available to the rest of the market, thus allowing them to either gain higher profit margins on the product or to lower the price to remain competitive.

6.3.5 Road Map for Digital Storage in Mobile Content Players Content Formats: Lossy and Loss-less Formats The MP3 music format is a lossy compressed music format created to economize on scarce and relatively expensive digital storage. MP3 and similar compression technologies generally compress music files to about 10 percent of the capacity size of the original content, removing most of the music “information” in the compressed file. Once this information is removed, it cannot be recovered—hence the term lossy. MP3 lossy compression takes advantage of the fact that the human auditory system doesn’t notice certain types of signal degradation, but lossy compression can introduce artifacts that can be noticed by a keen ear, particularly in a quiet background. This format has become very popular for use in relatively unsophisticated sound systems and in noisy mobile applications. A 10,000 song MP3 player requires less than 40 GB of total storage capacity. With the continuing decreases in storage prices it will become more common for users to move from MP3 to less compressed and even non-lossy compressed music in their portable devices. Loss-less compression reduces the storage capacity of audio content without losing the original audio signal’s integrity. Thus an audio track compressed with loss-less compression can be decoded to its original uncompressed form without artifacts. A loss-less compressed music file is at most reduced to 50 percent of the size of the original music file. A 10,000 song personal stereo player with 50 percent loss-less compression requires about 140 GB of storage. Loss-less compressed music download

www.newnespress.com

156

Chapter 6

services have become available and there are many programs to support loss-less ripping of CDs. Table 6.3 compares digital storage and streaming bandwidth requirements for various formats of music and video content. The numbers are approximations since compression rates and resolutions may vary somewhat. It is clear that richer digital content requires much greater storage capacity and bandwidth. Ultra-HD is assumed to be compressed as much as HD. A physical distribution media for mobile devices such as a small form factor optical disc or a flash memory device in a removable reader could be used for loading loss-less content. We expect that with higher capacity storage the need to economize storage space using MP3 files will be less compelling, thus driving the higher end of the music player market to use higher capacity hard disk drives. Figure 6.14 shows projections for the storage capacity that could be bought by an OEM for about $60 out to 2010 for flash memory as well as various form factor hard disk drives. Also on this chart are shown the storage capacity requirements for a 10,000 track MP3 and loss-less compressed music formats. It is clear from this chart that the MP3 player market will probably be almost entirely flash memory dominated by 2010, but if higher resolution content is required, then a hard disk drive based device may make more sense from a price and capacity point of view. Table 6.3: Storage, Streaming Bandwidth and Estimated Power Requirements for Various Formats of Music and Video Content Format Music Formats MP3 Loss-less Compressed CD CD Quality DVD Audio Video Formats Format for iPod (MPEG-4) DVD MPEG 2 MPEG 4 SDTV Blu-ray/HD DVD HDTV Ultra-HDTV

www.newnespress.com

Bandwidth (MBps) ∼0.128 ∼0.700 min. 1.400 9.600 max. ∼0.750 11.080 ∼1.400 ∼8.000 36.550 ∼19.300 ∼295.000

Storage Capacity/Hour (GB) ∼0.576 ∼0.315 0.630 4.320 ∼0.337 2.700 ∼0.630 ∼2.000 3.750 ∼8.890 ∼133.000

Storage in Mobile Consumer Electronics Devices

157

250 48-mm HDD 34-mm HDD 27-mm HDD

200

Capacity in GB

Embedded Flash

150 10,000 Loss-less Compressed Songs

100

50 10,000 MP3 Songs

0 2004

2005

2006

2007

2008

2009

2010

Figure 6.14: 1 Disk/2 Head (Minimum) Capacity Versus Hard Disk Drive Form Factor and Flash Memory Projections for ∼$60 OEM.

The very large digital storage capacity that is available with small form factor hard disk drives (or larger capacity flash memory) will develop a new class of personal media player. This player will provide significantly higher resolution audio and video content on an appropriately sized screen or some external viewing device such as the personal projector technology described earlier in this section. Just as greater memory on personal computers led to new features and higher performance, so too these new consumer electronics products will provide higher resolution and loss-less content storage, providing a more refined user experience. These sophisticated media player products will provide ready access to photographs and music files, as well as video files. In addition to playing prerecorded content, these devices will support DVR technology to record video and other content from cable, satellite, broadcast, or the internet. They could also be portable repositories for a collection of family photographs and video, increasing the capacity requirements considerably. The only limit to the resolution requirements of these devices is the size of the display available. If these mobile devices are used to play on higher resolution displays or projected images as discussed earlier, then there is almost no limit to the storage requirements that such devices could have. The following examples show the sort of storage capacity required: •

A combination 20,000 4-megapixel photo, 10,000 MP3 song, 100 VGA movie player would need about 130 GB.

www.newnespress.com

158 •

Chapter 6 A combination 20,000 8-megapixel photo, 10,000 loss-less compressed song (CD quality), 100 DVD (MPEG-2) movie player would need about 597 GB.

6.4 Cameras and Camcorders Digital still cameras mostly use flash cards of various formats. Very high pixel count cameras used by professional photographers sometimes use removable 1-inch hard disk drives in a compact flash form factor, especially if photos are being stored in an uncompressed format. It is likely that photograph resolution for professional use will continue to increase (it still doesn’t match the resolution available with traditional silver halide film) and thus I believe there will continue to be a niche market for removable hard disk drives for still cameras. The camcorder market is relatively small today, averaging between 10–15 million units a year. This small market penetration is probably due to the fact that the $500 to $2,000 price tag that many of these units command is too high for a large number of users. These high prices are driven in part by the recorder’s very intricate recording mechanism. Hard disk drive based camcorders are starting to make a dent in the market. One such unit, the $1,300 JVC Everio, boasts a 30 GB hard disk drive capable of recording up to 7 hours of MPEG-2 DVD-quality video. The attraction of these devices is that they can store a lot of images, and data transfers of video files from the camera to a computer are as fast as other external disk drive file transfers. Flash card based camcorders make lower cost camcorders possible. As the price of NAND in the near future supports the storage of reasonable amounts of video on an affordable quantity of NAND this market segment will take off. Once the expensive tape transport or other mechanical media handlers are designed out of camcorders, the subsequent cost reductions are quite likely to reduce average camcorder prices below $100, opening up the market to many more users than can currently justify such a purchase. It is quite likely that the market for $100 camcorders is 10 or more times the size of the market for $500–$2,000 models.

6.4.1 Layout of a Digital Still Camera Figure 6.15 shows a block diagram for a digital still camera. Digital cameras serve many different market segments all the way from very inexpensive low resolution

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

159

User Interface Graphic Buttons

CCD/CMOS Sensors

Digital Camera Front End Signal Processor

SDRAM

Flash Memory

Lens AF Shutter

Motor Driver

SD/CF/SM Card Image Processor

Mono Headphone

USB

Clock LCD Controller

Audio Codec

Microphone

TV

Source Drivers Gate Drivers

Mono Speaker

TFT LCD Display

LED Backlight

Camera Flash

Multioutput Solutions

CCD/CMOS AF/Shutter

Core/Processor Memory/I/O

Boost Converter

Buck Converter

Amplifiers

LCD

LED

LDO

LCD Drivers

LED Drivers

Supervisor

Power Management

Charger

Battery

Protection

Gas Gauge

AC/DC Adapter Battery Management

LEGEND Processor Interface Amplifier Logic Power ADC/DAC Other

Figure 6.15: Digital Still Camera Block Diagram.15 models to expensive professional units with resolutions of greater than 20 megapixels. Over time, the resolution of cameras in the popular price point of $200 to $400 has increased, with a doubling in average megapixels in digital still cameras roughly every four years. Figure 6.16 shows a teardown of a digital still camera showing various labeled components. The digital camera uses an image sensor (CCD or CMOS) to convert light directly into a series of pixel values that represent the image. The major components of a digital still camera are: •

15

The CCD or CMOS image sensor converts photons of light into electrons at the photosensitive sites in the CCD or CMOS image sensors.

Courtesy of Texas Instruments.

www.newnespress.com

160

Chapter 6 Rohm #BH6411 Audio Interface

NEC #mPD168110 Motor Driver

Renesas #M50231 Seven-Channel Motor Driver

Matsushita #AN2010IA Vertical Driver Matsushita #MN39483 4.2-Mpixel CCD Image Sensor

Toko #TX15460L 3.3-Volt Video Line Driver

Fuji Electric #FA7709 Six-Channel Switching Regulator Analog Devices #AD9948 10-bit CCD Signal Processor with Timing Generator

Figure 6.16: Teardown Image of Digital Still Camera (Nikon 4800 Digital Still Camera).16 •

The front end processor filters, amplifies and digitizes the analog signal from the image sensor using high-speed analog to digital conversion.

•

The digital image processor handles industry-standard computationally intensive imaging, audio and video algorithms. It also controls the timing relationship of the vertical/horizontal reference signals and the pixel clock.

•

The various memory stores executing code and image data files.

•

The audio codec performs digital audio recording/playback under the control of the DSP.

•

The LCD controller receives digital images from the camera and images them on the LCD.

•

The power conversion electronics converts input power from AC or USB to charge the camera battery and control the power management functional blocks.

6.4.2 Layout of a Digital Video Camera Digital video cameras are often referred to as camcorders. These devices have used small form factor digital videotape for capturing the video images for many years and 16

Courtesy of Portelligent, Inc.—teardown.com.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

161

these are still the most common digital camcorders. However, camcorders that record digital video on optical discs, hard disk drives, and even flash memory are becoming more common and probably eventually will displace videotape. Figure 6.17 is a block diagram of a digital camcorder showing the basic design elements in a digital camcorder. Figure 6.18 shows a teardown of a digital video camera showing various labeled components. In this device the hard disk drive for video storage includes an accelerometer and appropriate software to detect when the camera is falling so the heads on the hard disk drive can be retracted in order to protect the drive from damage.

6.4.3 Storage Requirements for Digital Camcorders Many modern digital video recorders allow capture of digital still images as well as video images. Hard disk drive based digital camcorders allow direct file transfer of data to a computer. Since the disk drive in the camera appears as just another computer disk drive to the computer, it is attached using a USB, Firewire (IEEE 1394), or eSATA Camcorder Control Buttons

1394 Bluetooth

I/O Control

Image Processing

CCD

IrDA

System Control

Microphone Microcontroller

USB

LCD Controller

LCD

A/V Out Controller

Monitor

Storage Controller

External Memory Card Micro HDD

Memory Controller

Internal Flash

ROM

RAM

Figure 6.17: Block Diagram of a Digital Camcorder.17 17

After Lattice Semiconductor block diagram.

www.newnespress.com

162

Chapter 6 Hard drive holder

Hitachi Metals Pkg. - #H48C Die 1 - Support circuitry (not shown) Die 2 - Three-axis accelerometer

Murata #EMC-03MA/-03B Gyro velocity sensors

Figure 6.18: Teardown Image of JVC Hard Disk Drive Based Digital Camcorder, JVC GZ-MG20.18 connection. Optical disc digital camcorders let the user remove the optical disc and play it back immediately on a DVD player or computer. Optical media for these applications has moved to blue laser discs for professional video cameras and it is likely that the market segment of digital video cameras that use optical disc media will move to blue laser products over the next few years. Flash memory and hard disk drives show relatively steady increases in storage capacity while optical disc formats have remained static for many years. The resolution of video as well as optical content in camcorders will continue to grow for the next few years. Digital still image megapixel resolution has doubled about every four years. With the growth of high definition commercial video in the home I expect that for the next several years video resolution in camcorders will on average double about every four years as well. The result is that multiple market niches are developing for digital camcorders. The very high end may continue to use some digital tape and hard disk drives, an endangered mid-range may use optical discs, and a very popular and rugged low end (lower resolution and lower cost optics camcorders) could offer lower resolution video at a very affordable price, storing the images on removable flash memory media. 18

Courtesy of Portelligent, Inc.—teardown.com.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

163

With the increasing storage capacity for an affordable price that flash offers, it may be possible to make sub-$150 and sub-$100 digital camcorders that can serve untapped markets. In general, the price of digital camcorders should decrease even though video resolution will likely increase as the expensive tape systems that are common today are replaced by less complex storage media such as optical discs, hard disk drives, and flash memory.

6.4.4 Road Maps for Camcorder Digital Storage As the resolution of integrated or stand-alone cameras increases, the price drops, and more people start to share their content, greater digital storage capacity will be required along with greater bandwidth on networks to share this content. It is likely that cameras will be equipped with wireless capability, making it easy to transfer content to a nearby computer on a local network and possibly even to backup images and video to a wireless mobile storage device that you carry with you on your Personal Area Network (PAN). Developments in camera optics could also be in for major changes as once-stodgy optical technology is transformed. Researchers are developing new devices using negative index meta-materials to create more complex and compact optical paths, allowing even more powerful and compact imaging systems. Multiple wavelength recording (not just visible wavelengths) is possible in a device that can be easily incorporated into other devices or turned into a lightweight and compact stand-alone imaging device. This multi-wavelength imager would allow the capture of much more than just surface images around us (imaging combining visible light, infrared light, and even higher or lower frequencies). New very small laser projection technologies could help turn digital cameras into their own projectors for playing back their content. All of these capabilities will encourage people to take more images and store more of those images in more places. Increasing ubiquity of still and video cameras, GPS location devices, and other technologies could also lead to a new type of device discussed later in this section that allows the recording of events as they happen in a person’s life and storing and even sharing parts of the content with others. The visual image is one of the most powerful ways for humans to communicate with each other and to learn about the world around us. It is only natural that we should

www.newnespress.com

164

Chapter 6

quickly incorporate these new advances in image capture, processing, and presentation to enhance the human experience and capture it for future generations.

6.5 Mobile Phones There is no more ubiquitous portable consumer device than the cell phone. Over one billion of these versatile devices are shipping annually. With such a huge market there are naturally many market segments with many different capabilities and price points. For many people, particularly in the developing world, their first phone is a mobile phone. Thus mobile phones represent a very important development in creating personal and business networks. In this section we examine the different types of digital phones and the digital storage needs and requirements for these products.

6.5.1 Basic Layout of a Cell Phone A modern cell phone provides two-way communications using one of several cellular standards (GSM, CDMA, TDMA, etc.). It often integrates one of more of the following functions: •

Vibrating ringer.

•

Polyphonic ringer.

•

Touch screen.

•

Still and/or video camera.

•

Broadcast radio receiver (FM, AM).

•

MP3 player.

•

PDA functions.

•

Mobile TV capability.

•

Location-based services.

•

Digital storage and storage expansion.

•

PC connection.

Figure 6.19 shows a reference block diagram for a cell phone including digital camera imaging capability. Note that the more features there are on the phone, the more

www.newnespress.com

165

Storage in Mobile Consumer Electronics Devices Fast IrDA

Emulator Mobile Pod DDR

Trace NOR Analyzer Flash

NAND Flash

GPIO

Level Shift 1

Sub Camera

Display

Camera Module

TWL 4030

Antennas

Battery Charger

3G Adjustable DC/DC For PA

DTV1000

Audio Codec

Applications and Digital Base Band Processor

WiLinkTM TNETW1253

HS USB Transceiver

Data Voice

BluelinkTM 5.0 Solution

Flashing Control/Data Voice

Speaker Amplifier

Keypad Control

TV PAL/NTSC QVGA Color TFT Display

USB Car Kit Interface (Optional)

Touch Screen Controller

Level Shift 2

Serdes Display Driver

XGA Color TFT Display

Speaker

USB Protection

Battery

External Charger (Optional)

Temperature Sensor

Power Manager

GPS53C0

TCS Modem Chipset

On/Off Power Source

Optional

USB Port

LED Driver Gas Gauge

Speaker Analog Amplifier Switch

Hi-Fi Audio CODEC Microphone

MS/MMC/ SD/SDIO Card

Headphone Amplifier

LED Flash

DTV Supply

OLED Backlight

LCD Backlight

System Power

LED

12C for Vcore DVS

Applications Processor

External Speaker

Headphone

LEGEND Processor Interface Amplifier Logic

Power Management

Internal Speaker

Power ADC/DAC Other

Figure 6.19: Block Diagram of a Mobile Phone.19

19

Courtesy of Texas Instruments.

www.newnespress.com

166

Chapter 6

memory, battery power, and processing power will be required. In particular, for capturing or storing higher definition still and video images, larger amounts of digital storage are required. The RF section of the circuit block diagram includes antenna diversity, downconversion, up-conversion, frequency synthesis, RF power amplifier as well as power ramp control, and other features. The baseband section of the circuit block diagram includes the DSP, Voice Band Codec (VBC), memory, and user interface controls for devices such as keypads, LCD, ringer, and so on. Figure 6.20 shows a teardown of a Sanyo S750 phone. This is a typical mass-market camera cell phone. The teardown of the main circuit board identifies the memory on the board, which includes some NAND flash and some SRAM. Most phones use solid state memory. NOR is common for system program memory and NAND flash is common for small amounts of audio, still photo, and video capture or playback. A small number of cell phones including some built by Samsung have included 1-inch and even the smaller form factor 0.85-inch disk drives. These small disk drives provide larger capacity for an affordable price for music or video files played on the cell phone.

Figure 6.20: Teardown of Sanyo S750 Cell Phone.20 20

Courtesy of Portelligent, Inc.—teardown.com.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

167

6.5.2 Development of Cell Phone Technologies and the Vision of Convergence Devices Cell phones differentiate themselves by offering various features in addition to the basic digital cell phone services. Cell phones can include still and video cameras, MP3 players, mobile TV viewers, and PDA capability including office productivity programs and files. At the time of the writing of this book, very few multiple-function phones did a very good job of handling multiple functions in a cell phone with acceptable performance and battery life and that were easy to use. There is a case to be made that multiple-function devices will seldom work as well as single or limited purpose devices. My own multiple function Microsoft windows PDA phone does an amazing number of things but many of them don’t work that well, and with so many buttons and ways to turn functions on, I never know what I will be pulling out of my pocket—a phone, a camera, a document, or a recording device. In addition, many modern phones will easily call back a prior called number resulting in what are called “pocket calls,” where a phone in one’s pocket calls someone up and they get to listen to whatever you are doing or saying at the time. Talk about a lack of privacy! There is interest in creating one mobile device that can serve for all one’s communication, information access, and work needs. Such a device could make our lives much easier but we are probably a long way from that vision becoming a reality. Convergence devices with multiple capabilities need to be more stable and respond to something other than just touch to activate most functions—it is too easy to touch a button by mistake when it is in a pocket or a bag. Voice activated devices would probably be better, since, hopefully, one would not accidentally say the wrong function code word and activate the wrong function. Voice activation and voice recognition, if easy to use and reliable, would be a big step forward in such devices. Today you can buy “dictation” software for your PC from multiple vendors that will recognize thousands of words, and that will capture, edit, and print a document by voice. You can call a voice activated assistant on the phone for banking, travel, ordering, and many other activities. The speech recognition algorithms used in these applications are quite large, consuming thousands of MIPs of processor performance and hundreds of megabytes of memory. Note that Dragon Naturally Speaking Professional can take as much as 2.5 GB of storage capacity depending upon the number of speech files installed. With greater processing speed in mobile devices and higher storage capacity (combined with noise cancellation technologies) we will have the critical requirements for

www.newnespress.com

168

Chapter 6

implanting usable voice recognition into mobile consumer devices. In recent years, algorithms based on modern embedded speech recognition have become available on many cell phones. They allow phone dialing by name or by number using your voice, and often allow voice activation of other cell phone functions. The application understands how names, numbers, and other words sound in a particular language, and can match your utterance to a name, number, or command in the phone. Users have found this new functionality straightforward to use and easy to remember. Worldwide, there are more than 20 million phones that include these modern embedded speech applications. They work very well, calling the numbers listed by name from your phone book, allowing you to dial your phone by saying the phone number, and letting you look up a contact entry, launch a browser, start a game, and more. Many of these phones use low power ARM processors. A few companies, such as Voice-Signal and Advanced Recognition Technologies, are working on low-power voice recognition solutions. The speech recognition software is running on hardware similar to that of a decade ago (except that it is very power efficient and small), but the speech algorithms are almost state-of-the-art.21 It is an easy step to see voice recognition technology becoming ubiquitous on mobile consumer devices. The use of displays that have contrast characteristics more like paper in ambient light would help in viewing content. In 2006, Motorola introduced the 9-mm-thick Motofone, which targets emerging markets like India and Brazil. This product uses an electrophoretic display, or EPD—an ultrathin, low-power display often referred to as electronic paper. While the yield of these displays is not good at present and the resolution does not match that of other display technologies, it is very low power and can be seen easily with ambient light. As this technology improves it could point the way to future lower-power and easier-to-see display technologies for mobile devices such as mobile phones. Adding additional functions to phones makes these devices more popular and allows offering subscription services that can help retain customers from going to competitors and competing technologies. This development, combined with the astounding growth of mobile phones of various sorts throughout the world, will result in a significant amount of development and experimentation on using phones as an element in personal networking and device convergence.

21

Source: IQ Online, Volume 3, Number 5, 2005.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

169

Most cell phones today work on cellular networks for their communication and downloading functions, but designs are now starting to appear that allow the device to use Wi-Fi, Bluetooth, and other wireless networking technologies as well. This is perhaps the best example of a mobile convergence device since the phone can act as a local media and data center for the user and perhaps for other devices that the user has. Incorporating IP wireless networks also opens up the possibility of creating PANs that allow communication between various devices that a person carries on him or her. Such PANs also open up the possibility of moving some digital storage and other functions off of a single device while still allowing for coordination and synchronization of content when needed. I believe that this may be the greatest means for near term convergence technology. In Chapter 8, we will explore how a local storage network could be used to provide high capacity, low cost storage to such mobile devices without drawing on their power budgets. A viable storage and communications network and improving local power management and power availability will lead to amazing advances in mobile consumer products such as mobile phones in years to come.

6.5.3 Importance of Mobile Power Sources and Cell Phone Size Power sources for mobile devices are a key factor in the usefulness of the device. For this reason, mobile products should always be designed to incorporate as many powersaving modes as are practical. This includes power-saving modes for storage products particularly where hard disk drives are used, because when these are actively spinning they consume a fair amount of power. New mobile phone designs incorporating media player functionality and other capabilities are allocating roughly 10 percent of their power budget to storage needs which, depending on the battery technology, translates to an average power consumption of about 110 mW. New features such as the laser pocket projector from Novalux described in Section 6.3.3 combined with all the other possible functions that could be incorporated into mobile phones could require considerably more power. In that section we also covered expected developments of mobile power technology that will enable more of these mobile applications. Also, technologies to harvest ambient power or to use wireless recharging capabilities tied to specially designed furniture with this capability built-in will enable a richer and more satisfying consumer experience. In order to keep the mobile phone at an acceptable size, either more power needs to be available in small power sources, or these applications need to use less power.

www.newnespress.com

170

Chapter 6

6.6 Other Consumer Devices There are several other types of consumer devices that utilize digital storage to enable their operation. We shall briefly discuss some of these as well as a short section speculating about future products enabled by very large mobile digital storage (more on this in Chapter 9).

6.6.1 Fixed and Mobile Game Systems Game systems use memory to store games. Home-based gaming systems that use a lot of graphics for their operation often use CD or DVD disks for game loading. Some of these systems also use hard disk drives to store game graphical video content so that it can be accessed quickly. The hard disk drives used for these applications are usually not very high capacity, but must be as inexpensive as possible in order to meet the bill of materials cost requirements. Mobile games tend to use semiconductor memory for game storage. The graphical requirements for the smaller screens do not require large storage files as do home-based systems, and so lower capacity flash memory or other nonvolatile solid state memory can be used. Use of solid state memory in this application is wise since these mobile game systems often get rather harsh treatment by the children playing with them. Multiplayer games played over the internet could change some of these storage requirements, particularly for the home-based game systems. This could lead to a greater amount of storage on a network and especially on the internet that gamers access. It is likely that resolution requirements and digital storage capacity requirements will continue to increase as other consumer devices increase in their video resolution and thus their storage capacity demands.

6.6.2 Handheld Navigation Devices Handheld navigation devices run the gamut from inexpensive systems selling for less than $99 to more elaborate systems that can cost several hundred dollars. The latter navigation systems are often designed to be used for automobile or marine navigation as well as for handheld use while hiking and other outdoor activities. As GPS receivers get built in phones and other devices and the price and size of these capabilities decline, it is likely that, like DVRs, GPS will become a standard function that many consumer devices will have built-in to their mobile devices in the next several years. Stand-alone

www.newnespress.com

Storage in Mobile Consumer Electronics Devices

171

devices that are just GPS map systems may become rare just as most DVRs are now part of set-top boxes. Regarding digital storage requirements for such devices, the storage needs are similar to those discussed in Section 6.2 of this chapter (on automobile navigation and entertainment systems). As the map details become more precise, the digital storage requirements increase. Also, if there is a need for real-time data updates to maps, terrain, or road conditions, a rewritable device such as flash memory or a smaller hard disk drive may be desirable. Figure 6.2 in Section 6.2.2 shows a block diagram of a GPS receiver unit. This is the key to a GPS map system. Data stored on some storage medium, here an embedded flash or SmartCard, contains the map data to display the location.

6.6.3 Other Mobile Applications Don Norman speculated about a Personal Life Recorder (PLR) type of device in his 1992 book Turn Signals Are the Facial Expression of Automobiles. He theorized that these PLRs would start out as a device given to young children, called the “Teddy.” The Teddy would record all of a child’s personal life moments, and as the child matured, the data could be transferred to new devices that matched his or her maturity level. Projects such as “My Life Bits” from Microsoft are also exploring the requirements for such devices. Someday we may use such devices to share clips of our lives with our families or friends or to help recall past events and contacts. Combined with the capability of organizing and indexing the saved content, such devices could provide a very powerful personal database that could be accessed any time. In this book we shall refer to such devices as life-logs. Such devices are today laboratory concepts, but in practice one could capture and sample one’s life including audio, video, and GPS information linked together. The pieces are now available and all that remains is for a creative and imaginative company to find a market for such a product. In an age of social networking and cameras in every cell phone, such a function could become very popular. With volume and with technology development, such mobile recording devices could become quite affordable. Depending upon the resolution, these devices would require huge amounts of information and the support system to store and organize the total aggregate content in the home could be even greater; a terabyte in the pocket and a petabyte in the

www.newnespress.com

172

Chapter 6

home would not be out of the question. All together the proliferation of such devices could lead to the single biggest use of digital storage in the world, since there are a lot of people out there and thus a lot of people that may want to record their lives.

6.7 Chapter Summary •

Automobiles will use digital storage for entertainment and navigation purposes. These devices must be able to withstand very harsh environments and continue functioning. As the complexity of automobile navigation aids and automobile entertainment resolution increases, storage capacity requirements will increase. Automobiles will also be part of the home network, allowing them to obtain content over this network. This could significantly automate the acquisition of content and navigation information.

•

Music and video player technology continues to develop. For highly compressed content (such as MP3 music) flash memory will probably provide all the digital storage needed in the future. However for less compressed (and higher quality) audio as well as video there will develop a lower end market for flash based devices as well as a higher resolution/content market for small form factor hard disk drive systems.

•

The use of a larger screen on video players allows the use of 1.8-inch hard disk drives that could provide much larger capacity for the price than flash systems for several more years. Thus these small drives may not suffer the same fate that 1-inch disk drives used in MP3 players such as the mini-iPod suffered.

•

New power source and wireless power technologies could provide extended usage life for mobile devices and enable the use of technologies that require more power, such as small projector systems for projecting images and video. Such systems could show higher resolution content than is required for smaller screens and thus require larger digital storage capacity.

•

Digital still and video cameras are being incorporated into many other devices as well as continuing to exist as a stand-alone device category. Lower cost flash based video cameras will create a very popular market category. These products will require larger onboard storage as content resolution increases and they will also require larger storage capacities for off-loaded content—such as into home storage direct or network-attached storage devices.

www.newnespress.com

Storage in Mobile Consumer Electronics Devices •

173

Ultimately the growth in available capacity of mobile digital storage components will enable the creation of devices that can record a person’s life experiences as they happen. Such a life-log will require significant digital storage capacity in the device as well as in the home network.

www.newnespress.com

This page intentionally left blank

CHAPTE R 7

Integration of Storage in Consumer Devices Objectives in This Chapter •

Demonstrate the contribution of digital storage devices to the costs of consumer devices.

•

Show examples of basic consumer applications that are becoming standard features in consumer devices.

•

Explore trends in the development of greater intelligence in storage devices.

•

Show how digital storage devices can be matched to various applications.

•

Project trends leading to the incorporation of applications in digital storage devices and the economic and performance implications of these trends.

7.1 Introduction Consumer electronics devices are undergoing significant evolution. Smaller semiconductor line widths enable fewer chips. With fewer chips, consumer devices can be smaller or they can incorporate more than one application within a single device. The content used in a consumer device must be stored in some sort of memory, whether that be a temporary volatile memory cache or some form of nonvolatile memory. Digital storage is an essential ingredient in modern consumer electronics products. Storage devices of all sorts are allowing larger digital storage capacity at ever lower prices, so more content can be stored on a consumer product. At the same time, the cost of digital storage for many consumer devices is a significant percentage of the total Bill of Materials (BOM) cost. Thus cost reduction of many consumer devices will directly

www.newnespress.com

176

Chapter 7

target digital storage. This could result in pressure to lower the total cost of the storage device or, alternatively, it could involve changing the storage device so it plays a bigger role in the end product performance with an overall reduction in end product costs. The standards used to define commands understood by digital storage products are defined by various ANSI standards committees. The T13 committee defines the ATA specification used in most hard disk drives and solid state drives. New commands are continually being added to the ATA specification, allowing more complex commands that enable features such as media streaming in consumer electronics products. Standards are also being developed for solid state storage devices in order to boost performance of various consumer functions as well. As consumer functions become more standardized, features such as GPS, radios, and DVRs eventually can be implemented as firmware in a single chip or small number of chips with appropriate supporting external hardware. Digital storage hybrid products are appearing, such as hybrid hard drives with flash memory incorporated into the hard disk drive or on the motherboard. These flash caches speed boot times and reduce power use by aggregating writes onto the drive. Such flash cache injections into storage architectures are examples of how combining different elements from the consumer electronics digital storage hierarchy can result in a product with features neither storage device could have alone. These developments create opportunities to use digital storage in new and interesting ways in consumer devices. Clever design of storage devices in consumer products provides greater functionality and performance, lower overall end product costs, higher reliability, and lower product footprint. This chapter explores the development of storage into consumer products today as well as looking where storage integration concepts could develop in the future.

7.2 Storage Costs in Consumer Product Design The following tables are representative hardware component bill of materials costs as a percentage of the total hardware cost for several common consumer electronics devices. These devices include digital storage in their design in order to support the availability of content that makes these products useful. These examples are shown to give an appreciation of the percentage of the cost of the device that is due to the digital storage. As we shall see, digital storage represents a significant part of the total product cost. This will have implications on future design of digital consumer electronics devices as shown later in the chapter.

www.newnespress.com

Integration of Storage in Consumer Devices

177

Table 7.1 gives a typical percentage breakdown of the percentage of total product cost for each of the key components (a BOM cost) in a representative home media server. Peripheral components such as cables and external storage are not included since these costs can vary and often are aftermarket products. As can be seen in Table 7.1, digital storage is a significant percentage of the total system cost for a digital video recorder and a larger percentage than any other single component. In an automobile navigation and entertainment system the proportion of BOM costs represented by the hard disk drive is 30–35 percent. Just as in the case above for a home media server, peripheral components are not included since the cost of these products can vary and often are aftermarket products. Table 7.2 gives a typical percentage breakdown of total product cost for each of the key components (the BOM cost) for an MP3 player with minimal display capability. Table 7.1: Home Media Server Component Cost Percentage of Total System Cost1 Component

Percentage of Total System Cost (%)

Motherboard Hard Drive (160 GB) 802.11n Card Power Supply Other

40.4 41.9 5.8 2.2 9.7

Table 7.2: Portable Music Player Component Cost Percentage of Total System Cost

1

Component

Percentage of Total System Cost (%)

Display Internal Memory (Digital storage) Battery Power Management Electronics DSP Electronics Interface Electronics (Such as USB) Audio Codec Housing and Other Mechanical Parts

3–5 65–75 Assume single AAA (Not included) 1.5–3.5 4–6 0.5–1.5 1.5–3 3–6

Source: Apple TV Design Stresses Volume over Profits, According to iSuppli, June 2007.

www.newnespress.com

178

Chapter 7

Peripheral components such as headphones, speakers, microphone, removable memory, and protective case are not included since these costs can vary and often are aftermarket peripheral components. As can be seen in Table 7.2, digital storage provides the majority of the total product cost for a mobile music player. Table 7.3 gives a representative percentage breakdown of the total product cost for each of the key components (the BOM cost) in a typical multimedia portable player (i.e., a video-focused personal media player). Peripheral components such as headphones, speakers, microphone, removable memory, and protective case are not included since these costs can vary and often are aftermarket products. As can be seen in Table 7.3, digital storage is the majority of the total components cost for a portable multimedia player. Table 7.4 gives a representative percentage breakdown of the total product cost for each of the key components (the BOM cost) in a representative multimedia phone. While not as large a percentage of the total BOM cost as the other components, digital storage represents the single largest percentage cost item in the BOM. As can be seen from these tables, digital storage provides a significant portion of the total cost of many modern consumer devices. In many cases, it is the single most expensive component in the device if not the majority of the cost of the device. Since digital storage capacity is likely to continue growing to meet the need for higher resolution content, this is likely to continue to be true for the foreseeable future. It is likely that the cost of digital storage as a percentage of total cost will increase in the future rather than decrease. Table 7.3: Personal Media Player Hardware Component Cost Percentage2 Component Internal Memory (80 GB 1.8-inch Hard Disk Drive) Accessories 2.5-inch Color LCD PCB/Connectors/Housing Video Processor Battery MP3 Player Processor Power Management Other 2

Percentage of Total System Cost (%) 61 8 5 5 5 3 3 3 7

Source: Apple Insider, iPod Classic: The Last Hurrah for HDD-Based iPods, October 11, 2007.

www.newnespress.com

Integration of Storage in Consumer Devices

179

Table 7.4: Multimedia Mobile Phone Hardware Component Cost Percentage3 Component

Percentage of Total System Cost (%)

Internal Memory (NAND Flash) Video and Audio Player Electronics Wireless Electronics Telephone Electronics Other PCB Components Camera Module Display Battery Housing and Other Mechanical Parts

28.1 13.5 7.7 7.7 14.9 4.4 13.4 2.1 8.2

7.3 Development of Common Consumer Functions Mature applications become standardized and with increasing chip complexity the number of electronic chips required to implement those applications decreases. Eventually these functions might be combined on one or a very small number of chips. This section explores some important consumer functions that are becoming ubiquitous on consumer electronics devices. The emerging standard functions that will be examined here are •

Digital video recorders.

•

Digital cameras.

•

GPS location services.

•

Wired and wireless network connectivity.

7.3.1 DVR as a Standard Consumer Function Not too long after the appearance of the original DVRs in 1999, designers began to look at combining DVRs with other entertainment systems and emulating the record and playback functionality of DVRs for non-video applications. DVR functionality began to be integrated in some DVD player/recorder devices to provide a single box 3

Source: iSuppli, Apple iPhone to Generate 50% Margin, According to iSuppli’s Preliminary Analysis, January 18, 2007.

www.newnespress.com

180

Chapter 7

DVD player and recorder as well as digital video recorder. DVR functionality has also been directly built into some TVs. The DVD recorder could also be used to “archive” content from the DVR before it was erased. In general, the content recording functionality is becoming one of those iconoclastic functions that are now incorporated into many products. TV tuners and appropriate software such as Microsoft’s Home Media Center are used to make desktop and laptop PCs record and play back television programs as well. There are also digital radio audio recorders offering an audio equivalent of digital video recorders. Today the set-top box and other products that integrate DVR functionality into another product are much more common than stand-alone DVRs. As a function such as digital recording becomes more widespread, it starts to be incorporated into general purpose electronic ICs. The function can then be enabled by appropriate firmware (software that runs inside a device). This bundling of maturing functions into chip sets will enable such applications as DVR or audio recording on space restrained devices such as mobile phones with TV reception. Also, the empty space in static home DVR devices such as that shown in Figure 5.2 can be removed, allowing smaller set-top boxes or DVRs. This reduction of space makes it much easier to include DVR functionality into more consumer devices. Note that the height of the hard disk drive may restrict the reduction of thickness of the host device, but other dimensions can be reduced as needed.

7.3.2 Cameras as a Standard Consumer Function Digital still and video camera capabilities are increasingly built into other consumer devices such as cell phones, PDAs, PMPs, and even computer laptops. Unless zoom and high magnification are required, the optics to create a camera are now easy to implement in a restricted space such as in a cell phone. Likewise, the electronics to make a camera have been minimized and cost reduced, making it easy to put cameras almost anywhere. Today almost everyone carries a video and still camera in a cell phone, and many people use these applications regularly. As these cameras have begun to show up on cell phones and other products, people are using them to capture photographs wherever they go and share them. The increasing presence of digital still and video imaging capability in more and more hands is one of the drivers of the social networking and content sharing that has become so popular.

www.newnespress.com

Integration of Storage in Consumer Devices

181

In the future, cameras will be expected to be a part of many consumer devices and to be used for many purposes. This is a technology well down the road toward being a standard consumer product function. The growth in the number of cameras that people carry with them opens up some interesting possibilities for continuously recording a life, creating a life-log, or for using multiple images from cameras to create different perspectives of events and activities or even to recreate a three-dimensional image of these events or activities with appropriate software.

7.3.3 GPS Location Services as a Standard Consumer Function GPS capability is being turned into a volume business with inclusion of location-based services in cell phones. As a consequence, the cost of GPS is declining rapidly and the footprint of the electronics required to implement this functionality has decreased. Thus it is becoming easier to include location services as a part of a consumer electronics device both from the economic and design point of view. Location-based services allow potential service providers new ways to create revenue from location-based advertising and offer users the possibility to create peer-to-peer location services.

7.3.4 Network Connectivity as a Standard Consumer Function Many mobile and static consumer devices now include some way to connect directly with local networks. Indeed many static consumer applications such as media servers are built around basic network connectivity. For mobile devices, wireless connectivity using Wi-Fi and/or Bluetooth technology are the natural way to provide network connectivity. In cell phones, an additional path for connectivity is through proprietary phone networks. Network connectivity allows a consumer device to exchange content with other devices on the network or over the internet. It allows synchronization and backup of content on these devices as well as streaming content from one source to another. It will eventually enable the home storage pool as well as personal area storage networks as discussed in Chapter 8. Practical digital rights rules that are agreed to by most users, content owners, and providers, as well as privacy protection will be critical to the full realization of the potential of network connectivity.

www.newnespress.com

182

Chapter 7

7.4 Intelligence of Digital Storage in Consumer Electronics The capabilities built into storage devices such as hard disk drives and flash memory have been increasing. This is because semiconductor line-widths have declined allowing more electronic functions on a given chip. As a consequence, more functions are built into and enabled by the firmware in a digital storage device. As an example, the T13 committee ATA specification for hard disk drives now includes commands that enable multiple video and other richer streaming content so that recording, sharing, and playing of content can go on simultaneously.

7.4.1 Security Providers Data storage device companies and those that design the products that they are incorporated into have been working together to increase the security of storage devices in mobile products (such as laptop computers). The Trusted Computing Group4 is one of the key standards organizations working to make such products more secure. The influential storage and host device companies that are members of this standards effort have created and published a Trusted Storage Specification. One application of this provides encryption of the user data where all aspects of the data encryption, including key management, are done on the storage device (usually a hard disk drive). This standard covers any storage device that incorporates security and trust services running within the storage device. The first products released that are covered by this specification are hard disk drives, so we will look at data encryption in hard drives as an example of this security service. With this method the encryption key never leaves the storage device. The encryption key and the encryption firmware are stored on a non-user-accessible area on the hard disk drive. Unless the user accesses the data using a password that enables the encryption key he or she cannot get to the data. If the encrypted key stored in the nonuser-accessible area of the drive is intentionally written over by the user, then the data on the drive is not accessible. The data encryption function on the hard disk drive is carried out by a function in the storage device called a Security Provider (SP). Security providers can be used for various security and trust functions or commands besides encryption services. SPs are 4

See www.trustedcomputinggroup.org/home.

www.newnespress.com

Integration of Storage in Consumer Devices

183

located on partitioned regions of non-user-accessible storage on the storage device. The hidden memory in one of these hidden partitions can be assigned to an application or more likely a basic application command since the size of the SPs is not very large (20–30 kilobytes). Use of such application partitions combined with integration of application hardware functions on a storage device could be important elements in the development of more integrated consumer devices. If desired, such security providers would be a natural location for DRM functions. Implementation of this architecture is achieved through interface specifications such as T13 for ATA and T10 for SCSI as well as the Trusted Computing Group. Figure 7.1 shows the general architecture overview of the trusted storage security provider.

7.4.2 Object-Based Storage A key mode to increasing the intelligence and capability of consumer electronics devices is the use of metadata. Metadata is data about data. Metadata summarizes and makes more useful the original data. Traditional metadata is file headers and other information about dates of creation and modification of a file, the software that was used to create it, and who the person is that made it. More sophisticated metadata, such

Enterprise

Trusted Drive

TCG/T10/T13

Trusted

Receive

Container Commands

SCSI

Application (on the the Host) Host) (on

Send and

ATA or

ISV

Firmware/hardware Firmware/Hardware enhancements for Enhancements for Security and security and cryptography Cryptography

Firmware Hidden Storage Security Providers

SP

Controller Storage

Support

Trusted

Assign Hidden Memory to Applications

Figure 7.1: Architecture Overview of Trusted Storage Showing Security Providers.5 5

Courtesy of Trusted Computing Group.

www.newnespress.com

184

Chapter 7

as that used in the creation of movies and videos, includes information on location, who is in the images, timing information, and other data that will be useful in editing the content. In a digital storage device metadata plays a similar important role in telling the hard drive what sectors of data blocks must be combined to create an actual data file. At the present time, although this metadata may be stored on the storage device, the actual reconstruction of content and the use of the metadata to organize and use the content is done in the microprocessor of a host device. The software that is used to recreate the data file from the raw blocks of data using the file metadata is called a file system. If a storage device included its own version of a file system, then the storage device would be able to access the data at least in part without the assistance of a host file system. If such a capability were built into a digital storage device, then that device could do many things that currently are done on the connected host device. This would include functions such as searching for a file. If a storage device could automatically index the files that it contains (i.e., by creating searchable metadata), then the storage device could offload, augment, or potentially even replace host system functions such as searching for files and content. If indexing software can be created that can sort and categorize still and video images and even sort by people or locations from a user’s own database of information and experience, then the potential capability of finding and using content in the storage device itself could be awesome. A storage device that includes such a file system is called an object-based storage device and the files on the storage device are referred to as the objects. The SCSI specification has included support for object-based storage devices for several years but the ATA specification used in many consumer storage devices such as hard disk drives has not included object-based storage support. Figure 7.2 compares a traditional hostbased file system to an object-based storage device. Note that for the object-based storage case, the file system is divided between the host and the storage device. By including at least some of the file system in the storage device, the host system can be freed to perform other functions, providing faster host functionality and potentially much faster performance. Imagine if there are several storage devices in a system such as a hard disk drive array and if, when a search request was made for files with some particular data, the disk drives could independently look for the data—the potential speed up in the search could be significant. Object storage devices and host-enabled systems are still not common, although they are supported in various operating systems. I suspect that they will become much more

www.newnespress.com

Integration of Storage in Consumer Devices

CPU

185

CPU Applications

Applications

System Call Interface

System Call Interface

File System User Component

File System User Component

File System Storage Component Sector/LBA Interface Block I/O Manager

Storage Device

OSD Interface File System File System Storage Component File System Storage Component System Storage Component Block I/OFile Manager System Storage Component Block I/OFile Manager Storage Component Block I/O Manager Block I/O Manager Block I/O Manager

Storage Device Storage Device Storage Device Storage Device Storage Device

Figure 7.2: Comparison of Traditional File System Storage to Object-Based Storage.6 common in the next few years as electronics integration (e.g., more powerful microprocessors on the storage device electronics) makes such additional capability more cost-effective and as greater performance is desired from consumer and other applications using ever larger storage pools.

7.4.3 USB-Run Software Applications Flash memory devices are being created with greater intelligence and functional capability. In essence, a USB device is like a small portable computer as shown in Figure 7.3. Because a USB drive has its own CPU and a file system, it is capable of performing applications itself. The USB Flash Drive Alliance (UFDA), a consortium of leading USB flash drive manufacturers, supports USB smart drives, which allow users to run active programs from USB flash drives. USB smart drives provide users similar functionality to desktop applications run from a flash drive. As of the date of writing, lighter weight programs 6

Courtesy of Trusted Computing Group.

www.newnespress.com

186

Chapter 7

I/O Controller

CPU

RAM/ ROM

NAND Controller

FLASH Flash

Figure 7.3: Schematic of a USB Drive Showing Internal Components that Make This a Small Computer. such as games, instant messaging, and photo editors, browsers with personal setting and bookmarks, as well as encryption, can be run from such smart USB flash drives. It is likely that as the storage capacity and capability of USB flash drives increases and if appropriate alliances are made with major office productivity software companies such as Microsoft then these devices can be used for more generally useful product applications. UFDA members include Lexar, PNY Technologies, Inc., Samsung Semiconductor, Inc., Buffalo, Crucial Technology, Infineon Technologies, International Microsystems, Inc., Kingston Technology, and Peripheral Enhancement. SanDisk (and M-Systems, since acquired by SanDisk) developed their own competing application platform for USB based devices called U3. Microsoft has signed an agreement with SanDisk to develop applications that can run in the U3 environment. Flash memory devices can contain data that is private or proprietary. Thus as with hard disk drive storage, flash memory devices with built-in technologies to protect the contents from non-authorized access are showing up in the market. A particular example of such a device is a USB storage device with built-in encryption. Examples of USB flash memory products with security capability include: •

Kingston DataTraveler Elite—Privacy Edition offers 128-bit hardware-based AES encryption and password protection.

•

Trek Thumb Drive Swipe is a biometric device that uses a reader to read fingerprints to allow access to the USB drive contents.

www.newnespress.com

Integration of Storage in Consumer Devices

187

7.5 Matching Storage to Different Applications For various reasons, different types of digital storage devices are better matched for different sorts of applications. Table 7.5 compares hard disk drives and flash memory storage capacities available for various enterprise, computer, and consumer applications. The choice of a digital storage device for an application is the result of an engineering and economic trade-off. As described in Chapter 1, as the technical and economic requirements for an application become clear there develops clear demarcations between the appropriateness for different types of storage devices, and thus there develops a storage hierarchy for consumer electronics applications. This hierarchy is only a guide, since there can be great value in achieving some of advantages of various storage devices by combining them into some sort of hybrid storage device containing more than one storage medium. The use of flash memory cache on a hard disk drive is an example of such hybrid storage technology (it should be noted that an alternative approach supported by Intel and others is to put the write cache on the host motherboard).

7.6 The Convergence of Electronics—When the Storage Becomes the Device—Or Was It the Other Way Around? The integration of device functions into smaller and smaller electronic packages has been enabled by the shrinking of semiconductor line widths as well as improvements in grounding and signal isolation in complex chip designs. Integration of hard disk drive electronics has allowed the creation of circuit boards that do not occupy the entire back of the hard disk drive like older disk drive generations. Even the 1-inch disk drive has a circuit board that is smaller than the drive enclosure. With higher unit volumes, expected developments in electronic circuit densities, and clever design, it is reasonable to expect that in a few more product generations consumer application electronics as well as storage device electronics could fit on a circuit board no larger than the dimensions of the storage device. Likewise, with denser electronics, flash-based devices could also be built with tighter integration into applications. As consumer applications mature they become more standardized. Standardized functions can be implemented on a chip with the functions accessible through firmware. As more and more consumer applications can be built into a few chips it is a natural step to look at how the digital storage can be most effectively integrated to reduce the complexity of the devices, improve their performance, and lower their cost.

www.newnespress.com

188

Chapter 7

Table 7.5: Storage Device and Application Requirements (Hard Disk Drives and Flash Memory) Application

2006 Storage Capacity

2010 Storage Capacity

Long Term Capacity Potential

Flash Application

HDD Application

Enterprise Storage

HDD: Up to 300 GB (SCSI/ FC) SSD: Up to 10s of GBs HDD: Up to 750 GB

HDD: Up to 1 TB (SCSI/ FC) SSD: Up to 100s of GB

Unlimited (Rapid Growth for Compliance and Content)

Yes, for Very High Performance Applications

Yes, for Performance or Capacity Applications

HDD: Up to Several TB

Yes, USB Drives, Other Removable, Hybrid Hard Disk Drives

Yes, for Internal Mass Storage

Notebook Computer

HDD: Up to 200 GB SSD: Up to 32 GB

HDD: Up to 1 TB SSD: Up to 200 GB

Unlimited (Rapid Growth Due to Rich Applications and SW Growth) Unlimited (Rapid Growth Due to Rich Applications and SW Growth)

Yes, Internal Mass Storage for Most Notebooks

DVR/PVR

HDD: Up to 750 GB

HDD: Up to Several TB

Yes, USB Drives, Hybrid Hard Disk Drives, Some High Performance Low Capacity Ultralight Notebooks Maybe for Camcorder Input— Removable Device

MP3 Player

HDD: Up to 12 GB Flash: Up to 8 GB

HDD: Up to 60+ GB Flash: Up to 32 GB

Yes, Dominate Internal Mass Storage

Hard Disk Drives Will Eventually be Displaced in Pure MP3 Players

Desktop Computer

www.newnespress.com

Unlimited (HDTV and Future High Definition Formats, No Limit to Hours Desired) 40 GB Is Probably Enough for the Maximum Capacity of a Pure MP3 Player

Yes, Internal Mass Storage (Could be 3.5-inch or even 2.5-inch)

Integration of Storage in Consumer Devices

189

Table 7.5: Continued Application

2006 Storage Capacity

2010 Storage Capacity

Long Term Capacity Potential

Flash Application

HDD Application

Advanced AV Player (PMP)

HDD: Up to 80 GB Flash: Up to 8 GB

HDD: Up to Several Hundred GB Flash: Up to 32 GB

Unlimited if Resolution Limits Expand and DVR/PVR Included

Limited Use of Flash for this Market by 2010

Digital Camcorder

HDD: Up to 80 GB Flash: Up to 8 GB

HDD: Up to Several Hundred GB Flash: Up to 32 GB

Yes for Low End Products Selling Eventually for Less than $200

Digital Still Camera

HDD: Up to 12 GB Flash: Up to 8 GB HDD: Up to 12 GB Flash: Up to 8 GB

HDD: Up to 60+ GB Flash: Up to 32 GB HDD: Up to 60+ GB Flash: Up to 32 GB

Depending upon Desired Resolution, Flash Could Open up Low End of Market 100 million

Excellent

34 MBps (at Least)

Common Network Technology Expensive Today Unlicensed Band, Interference Network Connection when Plugged into Power. AV is AudioVideo Version Said to be ∼30 MBps with Optional Extensions Possible Using Out of Band Transmission, MOCA Standards, or HANA Standards

www.newnespress.com

Development of Home Network Storage and Home Storage Virtualization

205

twisted pair, or fiber optical connections, providing data rates stated to go from 400 MBps to 1,600 MBps. The Home Phone Networking Alliance (HPNA) is another option for networking using twisted pair Plain Old Telephone System (POTS) telephone connections used in most homes. HomePlug is another option offering networking through the power wiring in a home. Appendix B gives a list of these networking standards groups. Some of these initiatives have better prospects than others and it is quite likely in the highly scattered consumer electronics market that more than one networking option will survive. Besides Ethernet and WiFi, MOCA and HANA have the support of major cable companies and entertainment companies respectively. This could help in pushing for their adoption. There are also companies pushing networking over power lines. However the Home PNA option of networking over phone lines has not been very public for the last few years and may not survive. If a home network is to be used for streaming entertainment applications, it must support data rates for the various types of streaming content. See Table 8.3. It is possible to create non-real-time media applications that involve downloading for an interminable amount of time on a slow network. The downloaded data is first put into a caching storage device, and then played in real time from cache storage. IPTV may use an approach like this. The storage devices themselves operate at much higher data rates than the networking infrastructure. A 3.5-inch hard disk drive used in a DVR, for instance, commonly provides 50 MBps data rate. Thus the network connections are usually the bottleneck for entertainment content or data transfer throughout the home. In practice a T-100 Ethernet network rated at 100 MBps runs at about 10 MBps including the effects of network overhead. It is clear that it would be difficult to handle multiple high definition media streams while still handling home data. For instance if

Table 8.3: Data Rates for Home Media Streams Home Media Stream MP3 Audio DVD Audio DV25 Video DVD Video HDTV Video

Average Data Rate (MBps) 0.02 1.2 3.0 1.25 5.0

Comments Typical MP3 12 : 1 Compression Very High Fidelity Digital Audio Common Camcorder Format MPEG-2 Video Blu-ray Disk Maximum Data Rate

www.newnespress.com

206

Chapter 8

10 MBps data (1.25 MBps) is required to support home data transfer requirements, the T-100 network could only handle at most one HDTV video data stream. The network could handle this general data traffic and seven DVD quality video streams. Many DVR/PVR suppliers can handle up to seven simultaneous video streams and as mentioned in Chapter 5 they are being designed for up to 16 simultaneous data streams. A Gigabit Ethernet network would allow roughly 100 MBps after overhead, while the Multimedia over Cable (MoCA) coaxial networking scheme allows at least 34 MBps data rates (HANA allows 50–200 MBps). Network data rates of at least 30 MBps can support a 10 MBps general data transfer plus five HDTV streams and several digital music streams. A T-100 Ethernet network could support DVD quality (MPEG2) video formats on a MoCA or HANA coax network. However for HDTV content a HomePlug power network, a HomePNA phone network with extensions, or a Gigabit Ethernet network will be required. Note that the HomePlug network would also combine the network connection with the power plug in, adding some extra convenience for the consumer.

8.4 Push Versus Pull Market for Home Networks Most of the growth of home networking has been based on retail rather than a service sales model. Up to the present, most home networks are installed and maintained by the individual homeowner or in some cases where the homeowner can afford it, by a home network specialist. This retail growth was driven primarily by sharing an internet connection as well as sharing some hardware resources such as printers. Additional growth in home networking requires new approaches that make it desirable and easy to implement. There is an enormous potential for home networking (and also home network storage) if it can be tied to a service that requires it. An illustrative example of the service model is the sale of entertainment content by cable or satellite companies. These companies usually supply the hardware, including the connection to the content source, at a significant subsidy or for free, and then charge the consumer a regular monthly fee for the use of the content and the hardware. In the case of most cable providers, service (including installation of the leased equipment in the home) is done by the provider company or its contractor. When the service provider supplies the hardware, it increases the available market by making it easier for less technically savvy consumers to own and operate the additional services.

www.newnespress.com

Development of Home Network Storage and Home Storage Virtualization

207

The result is a significant incremental source of revenue to the service provider. Even more important is that service providers find that offering increased services such as DVR/PVR capability in their set-top boxes increases their retention of customers, especially the most lucrative customers that demand and buy these additional services. Current U.S. cable households are somewhat more than 73 million today with about 25 percent of these having digital cable. Cable modems supply the highest percentage of broadband internet connections (about 60 percent of all broadband connections are through cable). With these overwhelming statistics it is natural to project that coaxial cable could be used within the home for media and file sharing as well as being the predominate source of external home networking through cable modems.

8.5 Home Networks for Media Sharing There are several approaches to a working home network. While this will increase the options for networking the home, it will also make it difficult for networking equipment manufacturers to ramp up a common networking standard. Until a dominant home networking technology is established, prices and availability may restrict growth. An approach for home networking that would appeal to a service model from the cable providers is using the existing coaxial networks in the home. Given the growing importance of sharing digital entertainment in the home, for cable connected homes this could become a dominant path for home entertainment (and data) networking. Strategic mergers and alliances in the industry, such as Cisco’s purchase of Scientific Atlanta, indicate the possible merger of home entertainment and data networks. As shown by the MoCA alliance (consisting of Entropic Communications, Comcast, Echostar, Panasonic, Motorola, Cisco, Toshiba, and Radio Shack) it may be possible to build a network using the excess bandwidth available in existing coaxial cables in the home to transport digital content and data up to at least 34 MBps. This is more than three times faster than most existing home wire-based Ethernet networking technologies. The HANA alliance specification offers even higher potential data rates. Only with the introduction of Gigabit Ethernet connections in the home will there be another home networking technology that can rival that of a coaxial network. The natural extension for cable service providers is to use this out-of-band (i.e., not using the same bandwidth that the digital cable signal uses) networking technology for home digital entertainment streaming and file sharing within the home. This could be

www.newnespress.com

208

Chapter 8

used as a way to share current set-top box DVR/PVRs. If so, then DVR/PVRs would become network devices in the home and share their content through the cable network with other PVR/DVRs or with home media players. Figure 8.1 shows how multimedia content can be shared through the home using Entropic MoCA technology. As an example of such DVR-based home networking, Scientific-Atlanta has offered Multi-Room DVR built around the company’s Explorer 8300 MR-DVR set-top boxes. This entertainment network is designed to allow consumers to access recorded content on multiple television sets within the home through existing in-home wiring and using existing Explorer digital set-tops. Scientific Atlanta’s 8300 MR-DVR serves as an inhome media server (see Chapter 5) that can be simultaneously accessed by up to three other, non-DVR Explorer digital interactive set-top boxes in the home. Multi-room digital video recording is also at the heart of Motorola Inc.’s Home Media Architecture (HMA), which incorporates technology from Ucentric Systems Inc. and Entropic Systems Inc. Ucentric provides home media networking software from a Javabased open architecture that enables content recorded on a DCT6208 or DCT6412

Broadband Multimedia

Den

Kids Bedroom

Master Bedroom

Thin Client STB Cable

PC

Client STB Coax

Coax

802.11 Wireless AP

Media Center

Family Room

WMA/MP3 Audio Client Kitchen

Figure 8.1: MoCA Model for Home Media Network.

www.newnespress.com

Development of Home Network Storage and Home Storage Virtualization

209

digital video recorder set-top box to be accessed from any HMA-enabled DCT digital set-top box. Entropic Communications provided an IP over Coax technology developed for MoCA. Field tests of the MoCA technology reported on April 4, 2005 at the National Cable show demonstrated that at least 100 MBps data rate was delivered to 95 percent of the coaxial jacks tested. Multiple tests were performed in over 200 homes and multi-unit dwellings, validating that the MoCA technology meets the requirements set in the Market Requirements Document approved by the MoCA Board of Directors. Entropic Communications’ c.LINK technology, the core technology behind the emerging MoCA standard, was used in the field trials. Thus the stage is set for service model introductions of media sharing in the home using home coaxial networks. Time Warner Cable has performed beta testing of the Scientific Atlanta multi-room DVR service in several sample markets. No system-wide introductions of coaxial multimedia sharing networks have yet been rolled out but given the widespread presence of coaxial networks and the need for cable operators to differentiate themselves from satellite and other competitors, the development of coaxial home networking for multimedia home distribution as a cable service option appears compelling.

8.6 Home Networks for Personal Reference Data Backup Once coaxial or other home networks for media sharing become popular, it is only a matter of time before these networks are also used for other networking. General data sharing would use more generalized IP over coaxial technology. Home computers are becoming the repository for gigabytes of home personal reference data. This home reference data consists of digital photographs, digital videos, digital documents, personal financial data, and other information that is of great value to a family and in many cases cannot be replicated if lost. Unfortunately a lot of this data is stored on single computer hard disk drives that are never or seldom backed up, and if they are backed up, the backups are seldom removed from the home to enable disaster recovery in case of earthquake, flood, or fire. It is not hard for a typical family to accumulate tens and even hundreds of thousands of digital photographs and digital videos over the course of a single child’s life. This amount of data can add up to hundreds of gigabytes and even more as the image

www.newnespress.com

210

Chapter 8

resolution of still and moving digital records continues to increase. By 2010 there could easily be over 2 TB of such content in a connected home as shown in Table 8.4. After 2010 we assume that one person in the household records their waking hours with a certain video resolution and frequency for the next 5 years (creating a life-log). As a result by 2015 a home can easily have close to 100 TB of accumulated personal content. A personal petabyte of all types of personal content over the course of a person’s life is not out of the realm of possibility. How can these precious records be preserved? If they are stored on a single hard disk drive the risk is too high that years of irreplaceable records can be lost. The only way to preserve this family reference data is to back it up and have multiple copies of the records available. Although it is not now a big driver for home networks, we are very close to home backup being a crucial part of the home storage environment. Backup might be the

Table 8.4: Projections for Growth of a Typical Tech-Savvy Home to 2015 Annual Generation of Home Reference Data Year Photographs Average Storage/Picture (GB) Number of Pictures Per Year Annual Photograph Capacity (GB) Home Video GB/Hour Number of Hours Annual Video Capacity (GB) Documents Average Capacity/Document (GB) Number of Documents/Year Annual Document Capacity (GB) Personal Diary/PMA Sampling Rate (seconds between Samples) Sampling Time (seconds) Resolution (GB/s) Diary/PMA Accumulated Capacity (GB) Total Capacity Per Year (GB) Cumulative Capacity (GB)

www.newnespress.com

2004

2005

2006

2007

0.003 500.0 1.3

0.003 550.0 1.6

0.003 605.0 1.9

0.003 665.5 2.3

0.004 732.1 2.8

13.0 11.1 144.3

14.3 12.2 174.6

15.7 13.4 211.3

17.3 14.8 255.6

19.0 16.3 309.3

0.001 20.0 0.0

146

0.001 22.0 0.0

176 322

0.001 24.2 0.0

213 535

0.001 26.6 0.0

258 793

2008

0.001 29.3 0.0

312 1,105

Development of Home Network Storage and Home Storage Virtualization

211

primary driver or a consequence of some other function such as remote access and synchronization between devices. As in the case of home networking, in order for home backup to become ubiquitous and frequently used, it must be easy and it must be automatic. Currently home backup usually involves external USB or Firewire (IEEE 1394) boxes. The performance of these interfaces was examined in Chapter 5. Ethernet connected backup boxes are starting to appear from companies such as Buffalo Technology, Lacie, Netgear, Olixer, Fabrik, and Seagate, but these are often not easily used to back up multiple devices. Ultimately, home backup needs to be built into the home network and to carry on its function largely independent of the actions of the people in the home. This could be done using existing Ethernet home networking, HomePlug, HomePNA, or could use an existing home media network using IP over MoCA coaxial cable technology with an appropriate combination of agents on the host devices and storage management software running on the storage network.

2009

2010

2011

0.004 805.3 3.4

0.005 885.8 4.1

0.005 974.4 4.9

20.9 17.9 374.3

23.0 19.7 452.9

25.3 21.6 548.0

0.001 32.2 0.0

378 1,483

0.001 35.4 0.0

457 1,940

0.001 39.0 0.0

2012

2013

2014

0.006 1071.8 6.0

0.006 1179.0 7.2

0.007 1296.9 8.7

27.9 23.8 663.1

30.7 26.2 802.3

33.7 28.8 970.8

0.001 42.9 0.0

0.001 47.2 0.1

0.001 51.9 0.1

2015 0.007 1426.6 10.6 37.1 31.7 1174.6 0.001 57.1 0.1

300.0

255.0

216.8

184.2

156.6

30.0 0.002 3,154

33.0 0.003 6,122

36.3 0.005 11,883

39.9 0.007 23,068

43.9 0.010 44,778

12,693 25,130

24,047 49,177

45,964 95,141

3,707 5,646

6,791 12,437

www.newnespress.com

212

Chapter 8

8.7 Projections for Home Network Storage As shown in Figure 8.2, in 2007 homes typically have at least three different communication networks that are more or less isolated from one another. These can include a computer network, a cell phone or two connected to a cell phone network, and an entertainment network (cable or satellite). By 2017 these isolated networks should collapse into a single network that is easy to use, low cost, and automates the most vital functions of home storage. A home storage network enables all the digital devices with storage that come into association with the home network part of a virtualized storage pool with automatic data replication, backup (both on-site and off-site), remote access and sharing of content, organization and indexing of the content, and data synchronization between devices and locations. Figure 8.3 shows what such an integrated home network could look like. The growth of networks in the home drives the growth of network storage devices in the home. Figure 8.4 shows projections for the worldwide growth of network attached

Internet Internet

Wireless Laptop

Multimedia Cell Phone Wireless Router

Desktop Computer

External Storage

DVR/STB

AV Player

Cable or Satellite

Figure 8.2: 2007 Digital Home with Isolated Home Networks for Computer Internet Access, Communication, and Entertainment.

www.newnespress.com

Development of Home Network Storage and Home Storage Virtualization

Internet Internet

213

Internet Backup Service Automobile DAVE

Multimedia Cell Phone

Wireless Laptop Router/Switch

AV Player Desktop Computer

External Storage

DVR/STB Cable or Satellite and Internet

Figure 8.3: 2017 Digital Storage Pool for Content Aggregation, Organization, Data Protection, Remote Access, and Sharing.

Worldwide Network Home Storage Devices in Millions

25.00 20.00

Multiple Drive Single Drive

15.00 10.00 5.00 0.00 2004 2005 2006 2007 2008 2009 2010 2011 2012

Figure 8.4: Projected Growth of Worldwide Network Home Storage Devices (Single and Multiple Drive Units).

www.newnespress.com

214

Chapter 8

Figure 8.5: Home Network Storage Devices.2 storage devices in the home.3 The actual growth could be even steeper once the proper marketing spin is developed to sell the value of these enabling storage products. Media sharing in the home or file sharing and remote access could be those killer applications that finally push network storage over the tipping point to become a real mass market product. Figure 8.5 shows some examples of today’s network storage devices. These products range from one to four drives providing various combinations of data protection and performance.

8.8 Design of Network Storage Devices An external hard disk drive storage device is a relatively simple product containing one or more hard disk drives, mounting hardware to keep the drive in the hardware (even if it can be easily removed), an adapter board to bridge Ethernet, e-SATA, USB, or IEEE 1394 (Firewire) communications to the hard disk drive. If there are two or more hard disk drives in the box then various forms of RAID (redundant array of independent—originally inexpensive disks) can be used to enhance

2 3

Photos from Storage Vision and CES Shows, 2007. Coughlin Associates, 2007.

www.newnespress.com

Development of Home Network Storage and Home Storage Virtualization

215

the overall system performance (such as read speed) or to increase the redundancy and reliability of the storage systems. A RAID is a group of disk drives that appear to be a single storage device to the user with attributes that depend upon the RAID classification. For two disk drives the two levels of RAID that can be created are called Level 0 and Level 1. Level 0 strips the data between the two disk drives. Data stripping refers to the segmentation of logically sequential data, such as a single file, so that segments can be written to multiple hard disks. Level 1 is mirroring in which the same data is written on both disks to provide good data protection from simple single drive failure mechanisms. RAID 0 provides very fast performance by the storage system but no protection of the data via redundancy. RAID 1 provides 100 percent redundancy by writing the same data to the two disks. As the number of disks increases, the RAID classification possibilities increase. Following is a summary of RAID levels commonly used (this is close to the classification of the original RAID systems in 1987.4 RAID-0 • Provides no redundancy, since the data are written across multiple drives (so-called stripping). If one drive fails, all the data in the array will be lost. •

Provides higher data rates, since all drives are accessed in parallel.

RAID-1 • Provides data mirroring and thus high reliability. The same data is written or read on two (or more) drives. •

Faster reading, since the first drive to respond to a request will provide data, thus reducing latency.

•

The cost at least doubles for a given storage capacity.

•

Higher data reliability.

RAID-3 • One extra drive is added to store the parity data (error correction data). If one drive fails, the data can be recovered and the other drives will keep working until the failed one is replaced (of course, performance will suffer). 4

Patterson, D., Garth, G., and Katz, R. (1987). A Case for the Redundant Arrays of Inexpensive Disks (RAID). University of Berkeley, Report No. UCB/SCD/87/391.

www.newnespress.com

216

Chapter 8

•

High reliability (cheaper than mirroring in RAID-1).

•

Very high data rates. Data writing and reading occurs in parallel.

•

For a given capacity, fewer drives are needed than for RAID-1.

•

Controller may be more complex and expensive.

RAID-4 • RAID-4 can be described as block-wise striping with parity. The parity data is always written to one disk drive and thus there is a reliability concern as well as the potential for a great deal of contention for the parity drive during write operations. •

RAID-4 is seldom used in modern drive arrays.

RAID-5 • Data and parity information stripping across all drives. •

High reliability, high performance.

There are higher levels of RAID that mix more functions together but a more recent and probably more useful categorization of RAID is given below.5 Failure-Resistant Disk Systems (Meets Criteria 1 to 6 Minimum) 1. Protection against data loss and loss of access to data due to disk drive failure. 2. Reconstruction of failed drive content to a replacement drive. 3. Protection against data loss due to a write hole. 4. Protection against data loss due to host and host I/O bus failure. 5. Protection against data loss due to replaceable unit failure. 6. Replaceable unit monitoring and failure indication. Failure-Tolerant Disk Systems (Meets Criteria 1 to 15 Minimum) 7. Disk automatic swap and hot swap. 8. Protection against data loss due to cache failure. 9. Protection against data loss due to external power failure. 5

RAID Advisory Board (RAB) 1996.

www.newnespress.com

Development of Home Network Storage and Home Storage Virtualization

217

10. Protection against data loss due to a temperature out of operating range. 11. Replaceable unit and environmental failure warning. 12. Protection against loss of access to data due to device channel failure. 13. Protection against loss of access to data due to controller module failure. 14. Protection against loss of access to data due to cache failure. 15. Protection against loss of access to data due to power supply failure. Disaster-Tolerant Disk Systems (Meets Criteria 1 to 21 Minimum) 16. Protection against loss of access to data due to host and host I/O bus failure. 17. Protection against loss of access to data due to external power failure. 18. Protection against loss of access to data due to component replacement. 19. Protection against loss of data and loss of access to data due to multiple disk failure. 20. Protection against loss of access to data due to zone failure. 21. Long-distance protection against loss of data due to zone failure. A RAID can be organized on a host using a group of disks in an attached host device or with the file system located on the storage device as shown in Figure 8.6. These are often referred to in the enterprise storage world as software and hardware RAID, respectively. Note that because of the overhead in creating a RAID array, the total storage capacity available to the user is less than that of the sum of the individual drive storage capacities. Also note that in theory a RAID array can be made with any group of storage devices that are networked together including tape, optical, hard disk drives, and flash memory.

8.9 Advanced Home Storage Virtualization A RAID storage system is an example of storage virtualization. The physical storage devices (the disk drives) are hidden behind the abstraction of the RAID storage so it looks like a single storage device such as a very large disk drive. All the other features of the array such as redundancy to protect the data in case of a drive failure are hidden behind this abstraction.

www.newnespress.com

218

Chapter 8 Volume Manager with RAID Capability

Host System

Host System

Host I/O Controller RAID Capability

or

Disk Drive

Disk Drive

Disk Drive

Subsystem RAID Controller

Disk Drive

Disk Drive

Disk Drive

Figure 8.6: Two Approaches to Creating a Storage Array. Software or Host Created RAID Array (File System on the Host) to the Left or Subsystem Created RAID Array (File System on the External Subsystem) to the Right.

There is no setup required to use a Network Attached Storage (NAS) device. You just plug it in and it looks like a big hard drive to the network and to other computers and devices that access it. Such abstraction of physical devices to create virtual (or logical as they are often called since they are the digital entity that can be directly accessed) devices is called virtualization. In hardware terms, virtualization refers to a simplified representation of hardware resources to the user, usually to simplify hardware management (e.g., multiple devices are represented as if they are a single device to the user. Virtualization can be applied to processors, storage, and various other digital devices to hide the complexity of multiple physical devices from the average user. Storage virtualization refers to hiding (also known as aggregating or abstracting) the actual storage resources behind some simplified user interface. As storage in the home becomes more widespread and complicated, virtualization of the various storage resources in the home will provide a way to simplify the management of the data scattered among the various storage devices. Figure 8.7 offers a graphic illustration of how the home storage virtualization stack may appear given enough complexity in available devices.

www.newnespress.com

Development of Home Network Storage and Home Storage Virtualization

219

Figure 8.7: Home Storage Virtualization Stack.6 Following is a description of the terms used in Figure 8.7.

6

•

A file system is a hierarchical structure for the organization of directories and the files within them. A file system is typically organized as a tree structure from the main or root directory.

•

Volume management is software used to aggregate or subdivide physical storage assets into virtual volumes.

•

A Host Bus Adapter (HBA) is an interface between a server or workstation bus and the Fibre Channel network. Hardware-based RAIDs are often organized at the HBA level.

•

Logical Unit Numbers (LUNs) are a SCSI identifier for a logical unit within a storage target. LUN Masking is a mechanism for making only certain LUNs visible to an initiator.

•

Zoning is a function of network fabric switches that allows segregation of storage nodes by physical port or World Wide Name.

•

A Virtual LUN is a LUN that is created that actually identifies several RAID or physical storage devices as a single LUN.

•

RAID is a redundant array of independent disks that is usually used to provide data recovery via data reconstruction in the event of disk failure. A RAID Group is a group composed of one or more RAID arrays.

•

Physical storage is an actual storage device such as a hard disk drive.

Source: T. Clark. Storage Virtualization. Reading, MA: Addison Wesley, 2005.

www.newnespress.com

220

Chapter 8

Storage virtualization can occur at several levels. It can be virtualization or aggregation at the block or file level, or at the disk level. It can be done at the host, in the network, or in the storage device or appliance. It is assumed that home storage virtualization will occur in-band (that is in the electronic transport used for the data itself). Figure 8.8 shows the types of storage virtualization important for home network storage as well as the various places that this virtualization can be done. Consumers use digital storage for various functions and thus there are typically many copies of content on computers, A/V media players, home servers, backup storage devices, automobile entertainment systems, and other consumer products. Increasingly, storage devices in the home are becoming the repository of a family’s valuable digital content such as photographs and family videos. As the storage on these devices grows, their number in and around the home increases. As the traffic on the home network increases, there will develop a need to extend the storage virtualization concept to all the digital storage in the entire home. A whole home virtualized storage system should provide content protection, efficient use of network resources, content synchronization where it is needed, organization of the content so it can be effectively found and used as needed, and ease of use.

Storage Virtualization

Block Virtualization

Disk Virtualization

Host-Based Virtualization

File System Virtualization

Network-Based Virtualization

File Virtualization

Storage Device or Appliance Virtualization

Figure 8.8: Paths to Home Storage Virtualization.

www.newnespress.com

Development of Home Network Storage and Home Storage Virtualization

221

8.10 Home Network Storage and Content Sharing within the Home In Chapter 5 we discussed bandwidth requirements for a network supporting a home media center with standard definition or HD content. In the future we could expect these requirements to increase even further. Next generation TV content with Ultra-HD (standards activities are just being started on this technology) would boost bandwidth and storage requirements significantly. Even the proposal for Ultra-HD does not result in fully real life clarity and does not provide a three-dimensional experience. Clearly, the storage requirements to support moving content around in the home will increase dramatically. As the number of devices that are members of proposed integrated storage or a storage utility in the home increases (probably with varying bandwidth and storage capacity requirements), there will be multiple copies of content used both as backup and to synchronize content among devices that can use it. This further increases the demand for digital storage. In the home it appears that the greater the number of devices that use content in the home network, the greater the bandwidth and storage requirements. Bandwidth requirements may be mitigated by approaches such as network coding, which may work particularly well in a home network environment. The basic notion of network coding is to allow mixing of data at the network nodes. A receiver sees these data packets that contain a mixture of data from all the sources and deduces from them the messages that were originally intended for that receiver.7 We will not get into the interesting specifics of network coding here, but if available network bandwidth can be used more effectively, then more content will be shared between devices in a network. As more content is shared, this content must be stored on either end of the transmission, either as temporary or longer term digital storage. In a very real way the required digital storage capacity is the physical manifestation of the content. The amount of storage can vary depending upon compression, but the basic transfer and storage of information is critical to the use of the content and thus the value of the network. In the next chapter we shall look quasi-analytically at the storage requirements for sharing digital files in the home as well as social networking and other ways that people

7

See http://www.ifp.uiuc.edu/~koetter/NWC/index.html.

www.newnespress.com

222

Chapter 8

are using to find new value in sharing their content with others. As we shall see, the digital storage requirements could be staggering.

8.11 Privacy, Content Protection, and Sharing in Home Network Storage In order for content owners to be comfortable with sharing content through a home network (such as might be connected to a home media center), some form of Digital Rights Management (DRM) needs to be implemented. Various schemes have been created to ensure that content does not stray from the people who have purchased a legitimate right to that content. These schemes often suffer the fault of making the use of this content inconvenient for consumers. It is clear that the proper balance needs to be established for the rights of content owners and consumers in order for the full market potential of today’s technology to be realized. Several forms of DRM are being considered in home media centers. These include: •

The Digital Living Network Alliance (DLNA)—for static home devices (appears to also incorporate the older Coral Consortium).

•

Open Mobile Alliance (OMA)—for mobile devices.

•

MPEG-12—also for mobile devices.

•

Security Architecture For Intelligent Attachment (SAFIA)—content protection scheme for iVDR removable storage cartridges.

•

Secure Video Processor (SVP)—conditional access for DVRs and other devices.

On the other hand, some content owners such as EMI are offering music content with a non-DRM option, making it easier to transfer and share the files between the users’ devices in and around the home. EMI charges somewhat more to purchase a DRM free version than a DRM protected version of a music track. Personal and family content may be stored and shared from home networks. This content must be protected from intrusion and theft as well in order to protect the privacy of the users. Various ways can be used to protect this content from unauthorized access. Ways to protect the privacy of personal content include:

www.newnespress.com

Development of Home Network Storage and Home Storage Virtualization •

Encryption of the content to prevent access without the encryption key.

•

Firewalls to prevent unauthorized outside access.

•

Some form of digital rights management for personal content.

223

The last approach to protect personal content is interesting in an age where more and more people are creating their own blogs, podcasts, and personal video sharing (various forms of social networking) over the internet. As the amount of this content grows (my own belief is that by 2015 there will be far more personal than commercial content), and as people share more and more of this content, there must develop some way for individuals to control access to this content and in some cases to create some way for people to get compensation for the use of their content. The result could be a very large micro-market for content created by individuals. These creators could allow access and use of this content by others for some small transaction fee. Approaches such as the digital commons could become the basis of such a personal content DRM, allowing some free access and use with attribution while allowing quasi-commercial access as well. While much content will be shared between tight social groups such as families, undoubtedly some of this content will intentionally or unintentionally find its way to a larger audience, and some means must be in place to protect the rights of the original content creator, or if appropriate, to create some way to compensate the creator for commercial use. With such a large supply of content available, finding shared content that fits one’s needs will become increasingly challenging. It is likely that many people will turn to aggregators of such content to provide access to this material. Such aggregators could provide the mechanism for quasi-commercial exchanges and micro-payments for the use of content, splitting the payments between themselves and the content creators.

8.12 Chapter Summary •

Many options are available and being commercially developed for home networking including use of wired and wireless Ethernet as well as using existing home networks such as coaxial cable, phone, and power lines. It is likely that more than one of these networking approaches will become commonplace in a market as fragmented as home networking.

•

While home networking is mostly a retail market with installation by the user or professional home installation service, this market could also develop as an

www.newnespress.com

224

Chapter 8 additional offering by service providers such as cable companies or telcos as a means of satisfying customers’ total entertainment and home networking needs and so to create a more sticky service offering in order to retain customer loyalty.

•

The largest general uses of home networks are for media sharing and home reference data backup.

•

By 2015, with the growth of personal reference data and with the development of life-logs to record a life as it happens, accumulated personal data in the home will greatly exceed commercial data and in general personal data will be much larger than commercial data.

•

We present a projection for the growth of network storage devices in the home with one or several hard disk drives and we discuss the design of multiple disk drive home network devices including RAID architectures.

www.newnespress.com

CHAPTE R 9

The Future of Home Digital Storage Objectives in This Chapter •

Estimate digital storage capacity demand created by personal content sharing in the home and across the internet.

•

Present the advantages and disadvantages of convergence versus single purpose devices and the implications for digital storage.

•

Study the impact of downloading on physical content distribution.

•

Look at the use of a life-log recording of ongoing life experiences to create a personal memory assistant and what that entails.

•

Explore the concept of a distributed managed home storage pool, the home storage utility.

•

Find out what new sorts of devices and business opportunities may be enabled by extensive personal life databases created by life-logs.

•

Show estimates for the growth of digital storage device units as well as storage capacities and the requirements for these applications in the future.

•

Examine the implications of digital content as our cultural legacy and history for future generations.

9.1 Digital Storage Requirements for Home Data Sharing and Social Networking In the home, as in other networks, the types of connections that can occur between devices that share content can be complex. These communications can be point-to-point

www.newnespress.com

226

Chapter 9

or multicast (going from one source to several receivers). There can exist short term and longer term associations between two or more devices in the handling of digital content. We shall create a simplified model to estimate the storage demands created by the sharing of content between multiple devices within a single home and between multiple homes connected through a network, such as the internet. This calculation will start with an estimate of single device storage requirements for retention and sharing of content and move on to cover the case of content sharing within the home. Finally, we shall extend this model to estimate required storage capacity for the case of content sharing between multiple homes and various other long and short term subnetworks that are part of a universal network such as the internet.

9.1.1 Storage Capacity Requirements for Single Use Devices In order to create an estimate of how much digital storage capacity is required for the home and ultimately for data sharing over the internet, let us start with individual device storage requirements. This analysis will be ad hoc but the assumptions used for the calculations as well as the formula will be included in order to allow modification of the analysis to fit other assumptions. The described case is used as an example only. The actual numbers that should be substituted in the formulas to calculate the storage capacity for short term cache and long term retention for a single use device may differ from these numbers. We shall assume a single purpose device in these calculations and we shall assume that the content will be completely downloaded before it is used. The time t is composed of two subunits of time. One is the time for the actual download td and the other is for using the content tu. If we multiply the data rate of content delivery D to the device by the download time, we get the resulting storage capacity needed for the download of content C (equation 9.1). C = Dtd

(9.1)

If content is to be used, the minimum time it would remain on the device is the total time t = td + tu. Thus the minimum storage capacity needed for the total time t is C. Depending upon the richness of the content and its intended use, tu may be longer or shorter than td.

www.newnespress.com

The Future of Home Digital Storage

227

As an example, if we have a connection between the single purpose device and some outside source of content that provides an available data rate to the device of 1,000 MBps (1 GB Ethernet at ∼100 MBps) and the resulting content has a capacity of 375 MB, then it will take about 3.75 s to download the content. If the content is then used immediately, playing out at a data rate of 1.25 MBps (standard definition video streaming rate), then it will take 300 s (5 min) to play out this content. Thus the total time for download and playback takes about 304 s (assuming that we don’t start playing until this download is completed). With these assumptions, we require the single use device to have a data storage capacity of at least 375 MB for 304 s. This is temporary cache storage which will be recycled for other uses unless the user decides to keep this content for a longer time. If we assume that two percent of the content downloaded is kept and that the device is used as described for two hours per day, then the retained content for this single use device over the course of a day is (7,200 s) × (375 MB/304 s) × 0.02 = 178 MB. Over the course of a year this single use device would be responsible for the accumulation of 65 GB of content. We can carry out the same sort of calculation for different sorts of content as shown in Table 9.1, assuming the same usage and channel data rate. In the table we assume typical MP3 tracks of 3 min and 5 min duration videos of standard definition and HD content. Devices that are used for personal content retention such as digital still and video cameras will have much higher retention rates (approaching 100 percent), but the daily and annual usage currently do not approach the two hours per day calculated here. If people in a home start to create life-logs with actual recording of their daily waking life, then the storage requirements will be much higher with a probable content retention near 100 percent. Table 9.1: Comparison of Stored Object Size and Annual Accumulation for a Single Use Device with Various Types of Digital Content Content

MP3 Audio Track SDTV Video HDTV Video

Playback Data Rate (MBps)

Content Size (GB)

td (Seconds)

tu (Seconds)

Annual Data Retention (%)

Annual Data Accumulation (GB)

0.02

0.0036

0.04

180

0.2

1.1

1.25 5

0.375 1.5

3.75 15.00

300 300

0.2 0.2

64.9 250.3

www.newnespress.com

228

Chapter 9

Many devices perform multiple functions and so their storage requirements will be different from those calculated here—probably higher if they combine video with audio content. Also, the long term content storage may not be in the device itself if it is part of a storage network. Long term retention in this case could be within a network attached storage device. Longer term retention could also be on a direct attached storage device that is not part of a network. In general the calculations in this section are meant to indicate that the desire to retain even a small amount of content after it is initially acquired (in this case two percent) will result in significant long term storage capacity accumulated demand.

9.1.2 A Model Home for Data Sharing In an actual home we can assume that there are a number of effective single use devices, even if some of the actual devices have multiple uses. The number of such single use devices requiring either short term cache storage and/or access to longer term storage capacity will probably increase with time as such devices become more and more a part of our everyday lives. It is a central thesis of this book that consumer devices with digital storage will have to communicate with each other in order to save personal and commercial content from loss, to effectively organize and use the content, and to share and synchronize content. In this section we examine the storage capacity implications of networked storage in a single home environment. As we estimated in the section above, the digital storage capacity requirements for a single use device in the home is given by Equation (9.1). In an actual home there may be a number of different devices with different storage requirements for content. We will create a model home with a number of single purpose devices requiring local content storage. In this home we assume two adults, two children, and a pet. The two adults share a desktop computer with a 500 GB hard drive and a laptop computer with a 160 GB hard drive and the two children share a desktop computer with a 300 GB hard drive. Between the two adults they have one smart phone with a digital still and video camera having 8 GB flash memory; one regular cell phone with MP3 music and MPEG4 video capability that has 64 MB internal flash memory and a removable 4 GB flash memory card; and one PDA (that is also used to play MP3 music and MPEG-4 video) with 8 GB storage capacity—probably flash memory. The adults use these two devices for their basic function (phone and personal organizer respectively) as well as the extended functions (each of these functions will be treated as a separate single purpose device).

www.newnespress.com

The Future of Home Digital Storage

229

The children each have a cell phone with a digital still and video camera with 32 MB internal storage and a 1 GB removable flash card as well as a mobile AV player that plays MP3 audio and MPEG-4 video, each having 4 GB storage capacity. In addition, the family has two digital still cameras with 6 megapixel resolution with 2 GB flash memory cards used in each. They also have a video camera with a 160 GB internal storage capacity (could be hard disk drive or tape). The family has two direct attached external storage devices with 750 GB and 500 GB capacity used to back up the two desktop computers and to contain some personal content, a 200 GB external storage device for backing up the laptop computer, and a 1,500 GB network storage device for secondary backup of the other storage devices as well as content sharing. The family also has a TV attached to a cable or satellite network with a DVR having 80 GB storage capacity. The devices and the assumed storage capacity in this model home are summarized below: •

Desktop computer with 500 GB hard drive.

•

Desktop computer with 300 GB hard drive.

•

Laptop computer with 160 GB hard drive.

•

Smartphone with 8 GB flash memory.

•

Regular cell phone with 64 MB internal and 4 GB removable flash memory.

•

Two regular cell phones with 32 MB internal and 1 GB removable flash memory.

•

Two mobile AV players with 4 GB internal memory.

•

Two digital cameras with 2 GB removable flash memory.

•

Digital video camera with 160 GB memory (removable tape or hard disk drive).

•

One external direct attached storage device with 750 GB storage capacity.

•

One external direct attached storage device with 500 GB storage capacity.

•

One external direct attached storage device with 200 GB storage capacity.

•

Network attached storage device with 1,500 GB hard disk drive storage capacity.

•

TV with DVR having an 80 GB hard disk drive.

www.newnespress.com

230

Chapter 9

The total raw storage capacity on these devices in and around the home is about 4,176 GB or about 4.2 TB spread across 17 devices including multiple use devices. This raw capacity does not include storage capacity on spare memory cards or commercial content on CDs or DVDs (including blue laser formats). The home network assumed in this model is an Ethernet T100 network (about 10 MBps data rate) and we assume that this network does not include the advanced integrated storage network or home storage utility concepts outlined in Chapter 8. This model reflects the common state of the art for today’s home storage networks. In enterprise storage systems 50 percent total storage utilization is pretty good, although virtualization could eventually bring this utilization to about 70 percent. In the case of home storage device utilization, this varies depending upon the device as well as personal practices and home economics. Content capture device storage is liable to be emptied regularly into longer term storage devices so devices such as digital still and video cameras, including those on mobile phones, are liable to run at a low average utilization rate (let’s say 10 percent). Mobile AV players, including those on the PDA and smart phone, are added to and subtracted from on a regular basis, and we will assume they run at 60 percent storage utilization and the content on these devices is primarily a copy of content on one or more of the personal computers and/or the network storage device. The personal computers will be assumed to run at 70 percent of total capacity. The direct attached external storage devices are used for backup of the computers that they are attached to as well as containing some unique personal content. We will assume that they run at 70 percent capacity. Likewise, the network attached storage device might contain an additional backup copy of personal content from the computers and DAS devices as well as some historical personal content that can be shared between the computers and other devices on the network. We will also assume 73 percent utilization for the network storage device. Because the DVR is constantly recording and filling its drive (while overwriting old content) we assume that this is 100 percent utilization. Table 9.2 shows the assumptions in this home model including content on the network that is commercial (such as purchased software, images, music files, and video files as well as personal content such as still and video images created by family members), computer files (e.g., text documents, spreadsheets, presentations, etc.), and any other content that originates from the home or family members. The model shows raw storage capacity, net capacity taking into account utilization, first instance commercial capacity, first instance personal capacity, and copy capacity (both commercial and personal).

www.newnespress.com

Table 9.2: Summary of Assumptions About Model Home Storage (Storage Capacities in Gigabytes) Device

Raw Capacity

Utilization (%)

Net Capacity

First Instance Commercial Capacity

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 —

500.00 300.00 160.00 8.00 4.06 1.03 1.03 4.00 4.00 2.00 2.00 160.00 750.00 500.00 200.00 1,500.00 80.00 4,176.13

70 70 70 70 60 20 20 60 60 10 10 10 70 70 70 73 10% —

350.00 210.00 112.00 5.60 2.44 0.21 0.21 2.40 2.40 0.20 0.20 16.00 525.00 350.00 140.00 1,087.50 80.00 2,884.15

170.00 110.00 40.00 1.00 — — — — — — — — — — — — 80.00 401.00

First Instance Personal Capacity

Copied Capacity

180.00 100.00 72.00 0.60 0.40 0.06 0.06 — — 0.20 0.20 16.00 175.00 140.00 28.00 393.00 — 1,105.51

— — — 4.00 2.04 0.15 0.15 2.40 2.40 — — — 350.00 210.00 112.00 695.00 — 1,378.14

The Future of Home Digital Storage 231

www.newnespress.com

Desktop Computer Desktop Computer Laptop Computer Smartphone Regular Mobile Phone Regular Mobile Phone Regular Mobile Phone Mobile AV Player Mobile AV Player Digital Still Camera Digital Still Camera Digital Video Camera DAS External Storage Device DAS External Storage Device DAS External Storage Device NAS Storage Device DVR Total

Device No.

232

Chapter 9

First instance content is the original copy of commercial or personal content which may be additionally copied on other devices for data protection or for local access when away from the network. In Table 9.2, commercial content such as DVDs (red and blue laser), CDs, and other content released on optical disks that cannot currently be legally downloaded into a home network are not included in this model and would likely add more than a terabyte of commercial content to that in Table 9.2. Note that this could change dramatically with the development of controlled access and other DRM standards that allow and support purchase of network rights for the home. With network rights, downloading of optical disc content as well as internet content to a home network would enable sharing of this content between various devices in and around the home. As stated elsewhere in this book, it is expected that total home content storage (including commercial and personal content) will be in the petabyte range for some homes by the mid-2010s. Also, as discussed in the section of this chapter on life-logs, the majority of this content could be generated by the members of the household. Note that in this model about 17.5 GB of new personal content resides on mobile devices that have not yet been incorporated into the home network and backup system. Improvements in virtualization and management of storage assets, such as those in these mobile content creation devices, could copy this content more readily onto the storage network which makes the content available to the household members and protects it from loss. Total first instance commercial content on the network totals 401 GB while first instance personal content totals 1,106 GB. Thus the total unique content that is available for sharing between devices on the home network is 1,507 GB.

9.1.3 Storage Capacity Requirements for Home Content Sharing Using Single Purpose Devices We shall tackle a calculation of the storage capacity requirements of a home storage network allowing sharing of content. We shall draw upon the single purpose device capacity equation and the network value formulas from Metcalfe’s and Reed’s laws. Bob Metcalfe, founder of 3Com Corporation and a major Ethernet designer, stated that “The value of a network increases exponentially with the number of nodes.”1 That is,

1

Answers.com/topic/Metcalfe’s-law.

www.newnespress.com

The Future of Home Digital Storage

233

where N is the number of nodes connected on the network, the value of being a member of this connected network increases as N 2. This is proportional to the total number of pairs of nodes in the network. This relationship is known as Metcalfe’s law. The actual equation for the total number of node pairs is given by Equation 9.2. N ( N −1) 2

(9.2)

where N is the number of participants. Since sharing of digital content generally occurs between individual content nodes as in Metcalfe’s law we can use this equation in our calculation of digital storage capacity required for data sharing. However, sharing might also occur between different groups of nodes (besides just pairs). In this case the total number of subgroups possible in the network follows Reed’s law as in Equation 9.3. Total number of subgroups possible = 2N − N − 1

(9.3)

We might equate the “value” of a network to the content shared across the network and the storage capacity used as the physical manifestation of this content, and thus a measure of content value. We will use these formulas for the number of potential interacting entities in a network to create equations that describe digital storage capacity requirements due to content sharing within a network. Within a home, this shared content is usually long term stored content, that is, data that is currently on the home storage network or that is added to the home storage network with the intention of long term storage. In a home storage environment, it is assumed that there is a master copy of the content somewhere in the home network and there may be short term copies elsewhere in devices permanently or temporarily attached to the network that are created by sharing content. We will make some estimates of the number of long term and short term (cache) storage in this calculation of home network storage estimates. At any point in time one or more of the devices in the household will be connected to the home network. Of these connected devices one or more of them might be sharing content with one or more of the other devices. For each of the home devices (preferably single purpose devices in this model even if the device can have multiple uses) there is a probability of participation at a point in time pi( t), where i is the number designating the device and t is the time. For instance, in the model household example i = 1 would

www.newnespress.com

234

Chapter 9

be the first desktop computer in whatever single-use mode it is in at the time, while i = 8 would be an AV player. Object i will be expected to be uploading or downloading content during some period of time ti. Due to the creation of both short term and long term copies of content during content sharing at a point in time, digital storage requirements could follow a form like Equation 9.4, if content is only shared between node pairs, or Equation 9.5, if content can be shared between other size groups. The digital storage requirements will be m2 − m ⎞ S = C ⎛⎜ ⎟ ⎝ 2 ⎠

(9.4)

for Metcalfe’s form (number of node pairs), or S = C(2m − m − 1)

(9.5)

for Reed’s form (number of all node groups). For these equations, m = pN, p is a group participation factor reflecting the activity of the subgroups (for the case of node sharing in the home, this is a function of the pi(t) described above), N is the number of nodes, C is the average size of shared content, and S is the amount of storage created by sharing. Let’s assume that we have a device that will connect for 5 min with one or more other nodes in a home using a T100 Ethernet network operating at 10 MBps. The size of the shared content is then 3 GB. We have 10, 100, and 1,000 node networks in the home and we assume in each case a 0.2 participation factor of all the possible sharing combinations participating in this sharing. Our results are shown in Table 9.3 for the Metcalfe and Reed approaches to calculating the total content sharing required storage capacity.

Table 9.3: Calculation of Storage Capacity for Content Sharing in a Network (GB) Number of Nodes

Participation Factor

10 100 1,000

0.2 0.2 0.2

www.newnespress.com

Metcalfe

Reed

3 570 59,700

3 3,150,000 4.82 × 1060

The Future of Home Digital Storage

235

Table 9.4: Calculation of Storage Capacity for Content Sharing in a Network with Reduced Participation with Increasing Node Number (GB) Number of Nodes

Participation Factor

Metcalfe

Reed

10 100 1,000

0.2 0.05 0.007

3 30 63

3 78 360

We see that for a constant participation factor for the network nodes, storage capacity increases enormously with the size of the network. In particular, in the Reed form, the growth in calculated storage is staggering. One can only assume that as the network grows, the participation factor must decrease considerably. If we assume a shrinking participation factor of the possible sharing combinations, we get a result like Table 9.4. The result of the lowering participation factor is that for 10 nodes we have two interacting node groups, while for 100 and 1,000 nodes we have five and seven interacting node groups, respectively. The resulting digital storage requirements are much less than for Table 9.3. Again, the Reed case is considerably larger than for the equivalent Metcalfe case. Only some of the digital capacity shared during this transaction will be translated to long term storage, so the net accumulated storage will be less than what is calculated here. It is pretty clear though that there is great potential for large digital storage capacities to support content sharing even within the home. For instance, with five node pairs in a 100 node network participating for 5 min, using the Metcalfe equation, 30 GB are required and using Reed’s equation, this increases to 78 GB.

9.1.4 Extension of the Content Sharing Model to a Larger Network We can extend this content sharing calculation for required storage capacity to sharing over the internet to find some truly large storage capacities. In order to keep the storage requirements to more reasonable level, we have assumed a network participation factor that drops significantly as the network size increases. Figure 9.1 shows the number of interacting node groups as a function of the number of network nodes (up to 100 meganodes assumed). This number of interacting node groups is equal to the number of nodes multiplied by the participation factor.

www.newnespress.com

236

Chapter 9

The resulting digital storage requirements for sharing as a function of the number of total network nodes are given in Table 9.5. We make the same assumptions for this case as for the case above with a 3 GB capacity being shared. We can see that it is very easy to come up with hundreds or even thousands of gigabytes required to share 3 GB of capacity in a sharing environment. If the participation is larger than that assumed, and if it includes sharing beyond just network node pairs, then the potential storage capacity required can be very large. As in the home network sharing case, the long term retained

12 Interacting Nodes

10 8 6 4 2

10 0, 00 0 1, 00 0, 00 0 10 ,0 00 ,0 00 10 0, 00 0, 00 0

10 ,0 00

1, 00 0

10 0

10

0

Number of Nodes

Figure 9.1: Assumed Number of Interacting Nodes Versus Size of Network (Number of Nodes) for Internet Sharing Model. Table 9.5: Calculation of Storage Capacity for Content Sharing Between Interacting Nodes per Figure 9.1 (GB) Number of Nodes

Participation Factor

Metcalfe

Reed

10 100 1,000 10,000 100,000 1,000,000 10,000,000 100,000,000

0.2 0.05 0.007 0.0008 0.000085 0.000009 0.00000095 0.0000001

3 30 63 84 96 108 121 135

3 78 360 741 1,058 1,506 2,141 3,039

www.newnespress.com

The Future of Home Digital Storage

237

content will be less than the total potential storage capacity required, but over time can be quite substantial.

9.2 Integrated Multiple Purpose Devices Versus Dedicated Devices When you turn your television on it comes on and is ready to use. When you turn your computer on you have to select an application and go through several steps to get that application working. There is always a struggle between whether it is better to design a device that is built to do a single function and a device that can perform several different functions. The television is a hardwired single application device, while the computer can do several functions depending upon the software used. It is this capability of programming a multiple function device that gives it the power to do many different things. This can be advantageous in saving the space and expense of having many single function devices. For a mobile device the vision of having one device that serves many functions such as a phone, personal organizer, camera, and so on, is very attractive. The danger of a programmable device is that in trying to do multiple operations it may make it difficult and almost impossible to do a single operation well. For this reason there is a fundamental conflict between ease of use and programmable functions in a consumer device. Some devices have done a better job of easing this conflict than others. Successful devices suit the needs of many consumers and are easy for most people to figure out how to use. One example of a device that is easy for most consumers to use is the digital camera in mobile telephones. These cameras have evolved over several generations so that they are small and can take a relatively high resolution image (although without an optical zoom function for the most part because of the limited optics and component cost concerns for a cell phone). Digital cameras have become a standard function (like those described earlier in the book) that can be installed in mobile phones. The main issue with digital cameras in phones is how the camera is controlled. The location and use of controls can make it easier to use the camera or too easy to take photos or videos accidentally. The design of controls in consumer products does impact the design of products using digital storage and thus should be taken into account. Consumer products with ample digital storage are likely to have multiple functions and thus these issues are

www.newnespress.com

238

Chapter 9

most particularly true for these products. The objectives for user interfaces are given below: •

The user interface should be intuitive to use, that is, most people should understand how to use it just by looking at it without additional instruction.

•

The user interface should anticipate how a user will use the particular function in this device and be in the right location for that function.

•

A user function should be enabled when (and only when) the user intends to use that function.

•

The user interface should be simple and easy to see or perceive in some other way, for instance if the device can effectively use human speech it should be apparent how to talk to it and what it understands.

•

The interface should use the best interface for the application it is used for both in terms of what people are used to for that function and in terms of the fit of the interface to that function—for instance if the device is to act like or provide control to a DVD player or television it might bring up buttons that emulate those used for these devices.

•

If multiple functions for a device will conflict or make the user interface too complicated then the number of functions should be reduced (or hidden) so that the user does not become confused.

•

A multiple function device that works well for users is likely to be a compromised device, and a full feature set professional device is more likely to be a single function device.

By following rules such as these, designers can make multiple function consumer devices that use digital storage easier for consumers to use and thus easier for them to adopt into their lives. Also, there will be a market for multiple purpose devices that work well but don’t have the full feature set of more professional devices that focus on providing more advanced feature sets. Thus putting digital cameras in mobile phones does not destroy the market for more advanced single application digital cameras with zoom functions, advanced “f”-stop settings, and the other things that amateur or professional photographers require. Rather, such devices extend the applications for cameras for more professional users while the cameras in phones make it possible for anyone anywhere to take at least a recognizable photograph of something.

www.newnespress.com

The Future of Home Digital Storage

239

9.3 Physical Content Distribution Versus Downloads and Streaming Until the mid-2000s, physical content distribution, principally using various optical media, was the dominant mode to distribute most commercial content and personal content. Mobile USB storage devices have supplanted optical discs for most physical personal content distribution. Many people have several of these devices that they use to carry their own content as well as to copy content from other people. With the increased availability of high speed internet access, downloading of music, video, and other content has become very common. While this content distribution has often been done without a return to the content creators or distributors in the past, content owners and distributors have begun to offer options to customers that are attractive for legitimate content purchase and distribution online. The internet has also enabled other less formal distribution technologies such as social networking and sharing and copying music, photos, videos, and so on, using web sites such as YouTube and many others. People are also putting their personal content on remote sites to make this material available to others such as family members and friends. This sort of content is often available using passwords or other means to control access and protect personal privacy. Earlier in this chapter we explored the implications of sharing content outside of homes on digital storage requirements and found that social networking and other new content distribution technology will create great demand for digital storage. There may be a fundamental issue to how much content can be shared using the internet. The internet connection between users has only a finite bandwidth. The content that users can share in a set amount of time over the internet can be limited by this bandwidth. This may be a further stimulus for the development of intermediary online storage services that allow sharing of personal content with the only limit being the bandwidth and time available to the person accessing the content rather than having content limitations on both sides of the exchange. With very rich content such as raw uncompressed video, it may be that physical content distribution still makes the most sense. There is an old saw that a FedEx truck full of physical media such as DVDs has a lot more effective bandwidth than most internet connections and is likely to be less expensive as well. People may also want to customize their content for DVD players and other platforms, and physical content

www.newnespress.com

240

Chapter 9

distribution may give them the best way to do this, particularly with current DVD (red and blue laser) and optical player technology. It is likely that online distribution will exceed physical content distribution by the 2010 decade, but that will probably not eliminate physical content distribution. Optical disc rich format content distribution will likely continue to be the most assessable and cost-effective means for commercial content distribution. This also fits into the current distribution models, including online purchase of such content. Physical content, if content protection allows local working copies such as on a home media server, can provide an effective backup of that content. Also, the content can be restored to the media server from the original copies should the media server content become corrupted.

9.4 Personal Memory Assistants In earlier chapters we described a life-log that could record images, sounds, time, and location of experiences of a person during the course of his or her life. Such a life-log could be the basis of a device you can carry with you that can act as a Personal Memory Assistant (PMA). This device, combined with a capability of a home storage utility (described below) to organize, summarize, and synchronize the content from a life-log and a sophisticated image and voice recognition (and reconstruction) technology could create a device that can help you remember places and people that you have interacted with in the past. Such a device would truly be a prosthetic memory device and could be used to help those with failing memories as well as enhancing already good memories. This is a list of possible attributes for such as device: •

Storage of compressed images and sounds (and perhaps other senses as well) sampling interactions with people and locations.

•

Creation of a personal map of where you have been and when, combined with associations with these images and sounds.

•

The ability to search this representative database of images and sounds rapidly to find prior occurrences.

•

The ability to appropriately model the effects of age and time on people and places to allow more accurate identification of prior occurrences.

www.newnespress.com

The Future of Home Digital Storage

241

•

The ability to tie into online resources (wirelessly) to enhance the information on the person or place as required.

•

Rapid creation of an accurate and succinct summary of past associations and a means to share this content with the user (such as a voice that whispers in your ear telling you who the person is that you are seeing and when and where you saw them before).

•

An easy to use and unobtrusive interface as well as a size, shape, and weight making it easy to use anywhere.

A PMA would require a considerable digital storage capacity as well as sophisticated computer processing and software. The storage capacity required will probably exceed 10 TB to carry sample content and metadata representing a significant portion of your life and such storage capacity (as well as the other capabilities) in a mobile device probably will not be available until about 2015 or so. Figure 9.2 is a block diagram of such a device. Ultimately such a personal organized database of your life and experiences could be used as the basis of some interesting entertainment applications as well as giving someone else the ability to experience some of the things that made you who you are. There are obviously many privacy and other issues to be worked out to make such

Off-line processing Processing in in Home home storage Storage utility Utility

Digital Storage (>10 TB)

Personal Map of Experiences, Places, Places and Times

Experience Capture Hardware HW and and SW Software

Life Search Function

(Capture (capture metadata Metadata includes Includes Location locationand andTime) time

Life-Log Device

User Interface and Privacy privacy protection Protection

Wireless Background background Search search and Compilation compilation

Figure 9.2: Block Diagram of Personal Memory Assistant Showing Major Component Functions.

www.newnespress.com

242

Chapter 9

devices universally used, but it is my belief that the technology to make such a device is within our grasp.

9.5 Digital Storage in Everything A key point of this book is that digital storage is a key technology for the creation of modern consumer electronics devices. Digital storage is the physical manifestation of the commercial or personal content that the consumer device creates, displays, or interacts with. In this section we will explore the implications of the consumer electronics digital storage hierarchy on the future of consumer electronics as well as management of the pockets of digital storage in all the consumer devices in the home and the threat of format obsolescence on the longevity of consumer content, particularly personal content. As electronic integration increases, it will be possible to integrate more and more standardized consumer applications and application functions on the actual digital storage device. This would save on the cost of the application, increase the application performance, and probably increase total device reliability for most applications as discussed in an earlier chapter. Today this would be easiest to implement on the circuit board of a device such as a hard disk drive but as electronics shrink further, these applications could also be incorporated into the circuits of solid state storage devices such as flash memory. This intimate tie-in of digital storage and the consumer applications should also incorporate standards making communication and management of the content between various consumer devices in the home much easier. As digital storage of various sorts is required in all consumer devices, making sure that this content is managed becomes a greater and greater problem. The home storage utility discussed below will eventually take care of this coordination. Until an overall home storage management application becomes available, people will use an ad hoc collection of applications that synchronize and back up at least some of the content. This is particularly important for content capture devices such as cameras, since that content will not be available anywhere else if the original source is lost or broken, unless it is actively managed. Likewise, there is an issue with the obsolescence of consumer devices. Will the content on these devices be moved to the replacement device and if there is original personal content on these devices, is it moved to a location where it can be backed up, protected,

www.newnespress.com

The Future of Home Digital Storage

243

and organized? This is a growing issue as consumers have an increasing number of older devices in their garages and closets. Efficient ways need to be created and standardized to make movement of applications and content to new devices easy and effective. If all devices could be part of a network, in many cases a wireless network with appropriate management software for moving and protecting content as well as converting to a new device, this management will be much easier. Today there are no such clear standards to make these functions easy and so users must use semi-manual ad hoc approaches to move content to new devices. Format obsolescence is a clear threat to the preservation of digital content and the applications that use this content. In fact the content formats themselves are subject to format obsolescence, as more efficient compression and other technologies are developed. Over time the older formats tend to lose support, making it likely that some older format content could be lost with time. If there is a large volume of older format content in many places, then it can be a daunting task to move this content to the new format. The Storage Networking Industry Association (SNIA) has begun an effort to deal with the issues around creating long term archives. The SNIA’s 100 Year Archive Task Force is operated by the SNIA’s Data Management Forum as a global, multi-agency group working to define best practices and storage standards for long term digital information retention. Although these efforts are focused on long term enterprise data retention, hopefully some of these efforts could also benefit consumers with stored content. In a sense the only way to really archive digital content for the long term is to actively manage this content, moving it to new formats as soon as they become established. This is one of the functions that an integrated home storage network and ultimately a home storage pool should provide.

9.6 Home Storage Utility—When All Storage Devices Are Coordinated In the home there are many devices that incorporate digital storage in their design. If these devices are connected to a network, then they can be addressed by devices on that network. In a home environment, virtualization of storage could reach an even more atomic level to address issues with protecting, sharing, and organizing home digital

www.newnespress.com

244

Chapter 9

content. These networks should back up and protect user content wherever it is. We may eventually see the development of what I call a home storage pool in which all devices connected to the network either continuously or intermittently (such as mobile devices and automobiles) would be virtualized into a single managed storage system referred to as a home storage utility. The home storage utility should provide the following basic functions: •

Content backup in the home including deduplication to make better use of network and storage resources.

•

Content backup outside the home (to provide home disaster recovery).

•

Content sharing in and around the home with optimal use of network resources.

•

Indexing and organizing home content so that it can be found and used when needed (particularly there needs to be a home automated metadata generation facility).

•

Synchronization of content as needed (e.g., delivering music and video to A/V players).

•

General management and control of storage and network resources (e.g., it should automatically identify and troubleshoot problems, letting the owner know how to solve problems in a way they can understand).

This is a lot to ask of a home storage network and we are nowhere near this level of automated management of home storage resources today. The development of the home storage utility will probably occur in stages. We have identified six phases of home network evolution, each involving ever increasing levels of storage networking, larger content files, greater need for storage services, and eventually complete virtualization of all home storage resources. Figure 9.3 displays these phases in a graphic form.

Phase 0: Internet Sharing Networks In the Internet Sharing Network, a home network is created for the sole purpose of sharing an internet connection between two computers. This has been the major driver of home networking up to the present time. Since there is no true storage networking we are referring to this as Phase 0.

www.newnespress.com

The Future of Home Digital Storage

245

Internet Internet Sharing Sharing Networks Network File File and and Peripheral Peripheral Sharing Sharing Networks Network Backup Backup Networks Network Home Home Media Media Sharing Sharing Networks Network Integrated Integrated Home Home Storage Storage Networks Network Home Home Storage Storage Utilities Utility

Figure 9.3: Phases of Home Storage Network Evolution.

Phase 1: File and Peripheral Sharing Networks In a File Sharing Network, the connected computers on a home network share document and perhaps MP3 music files as well as peripherals such as scanners and printers.

Phase 2: Backup Networks Once a network is in place with storage attached, storage devices on the network can use the network to back up their files on other storage devices on the network. Often this would occur on a specialized network backup appliance.

Phase 3: Home Media Sharing Networks In a Home Media Center Network, large files such as those generated by video are shared in the home network with other computers or connected devices such as PVR/DVRs, STBs, or AV Players.

Phase 4: Integrated Home Storage Networks An Integrated Home Storage Network combines the above features together in a home to allow simultaneous internet connection sharing, file and peripheral sharing, large multimedia file sharing, and data backup. The Integrated Home Storage Network may also include some connection to fixed and mobile storage devices in the home such as those incorporated into PVR/DVR/STBs, mobile MP3 and AV players, cell phones with

www.newnespress.com

246

Chapter 9

multimedia capability, automobile navigation and entertainment systems, and other mobile and fixed consumer devices using digital storage. As networking technology develops, it may eventually incorporate ways to more effectively utilize storage and network resources such as deduplication to avoid making multiple copies of the same data file or a coded network2 that allows combining multiple content into a packet that can be deconvolved by the target devices using data provided separately from each of the target devices to all the others. It may also be tied into a home security and sensor management system and so become part of a home automation solution. Phase 5: Home Storage Utilities This term refers to the virtualization of all mobile and fixed device storage in and around the home. This phase of home network storage requires standards based virtualization of the storage devices in all of the mobile and fixed consumer devices in the home. Virtualization of all the storage devices should allow easier DRM support, protection of user privacy, overall storage and content management and reduced user and system support. The user interacts with the utility to acquire and use any valid content. By the end of 2010 we expect to see the development of pervasive home storage networking technologies leading to a true home storage utility in the following decade. Note that these phases can co-exist and that some of them may not follow in succession. The growth and progression of these phases are driven by three major factors. First, home networks allow sharing of files and other resources and backup of home reference data similar to their uses in businesses. Second, home networks will allow sharing and distribution of entertainment content in the home. The third factor and the major motivator for the development of the last two phases is an increasing need to coordinate the various types of data and locations where that data may exist throughout the home. As multimedia and traditional files grow it becomes harder to find and organize this data. There is a great opportunity here to develop networked storage systems that can automate the organization of home content, make sure that it is where it is needed when it is needed and that it is protected from loss.

2

R. Ahlswede, N. Cai, S.-Y. R. Li, and R. W. Yeung. Distributed Source Coding for Satellite Communications. IEEE Transactions on Information Theory 45(4), 1111–1120, 1999.

www.newnespress.com

The Future of Home Digital Storage

247

9.7 Digital Storage in Future Consumer Electronics As discussed earlier, large and dispersed digital storage capacity in consumer devices will enable many new capabilities as well as drive people to develop technologies that can organize and effectively use such large and fragmented content. It is important to realize that the next 10 years will turn most of those who use consumer electronics into effective creators of digital content, and that the personal content created by these users will greatly exceed the amount of commercial content that will be stored in their homes and anywhere else for that matter. As personal content becomes the majority of digital content, this content will become a big factor in content use. If this personal content is organized and easy to access such as that in a PMA, then this content could provide material that could be used in new entertainment and communications products. A well organized personal memory database would provide material for new capabilities such as the following. •

Entertainment such as games or movies could incorporate personal experiences (such as people you have met or places that you have been) to make a very personalized and more gripping experience.

•

Multiple people using such a game or movie could share some of their own personal experiences to make a group experience that reflects in some way their shared experiences—whole new types of entertainment could be built upon these principles.

•

Users could create life mashups of their daily or longer term experiences that they could share with others.

•

Business opportunities will be created based upon sharing, combining, and creating personal stored experiences and this could well be a larger business sector than with pure commercial content.

•

A home automation system tied into elements of the personal database could better accommodate the people that live there or visit based upon previous stored experiences.

•

Automated systems could assist users in meeting their personal objectives and obligations based upon their personal database.

In addition to the content creation capabilities afforded by mobile high storage capacity (or access to large storage capacity) devices, the home storage pool or storage utility will play a very important role in the sharing of content. The creation of a unified

www.newnespress.com

248

Chapter 9

storage utility in the home should simplify content sharing and make the process more efficient. This will increase the capability to create and share content. As people share their created content on the internet the role of aggregating and organizing content will become ever more valuable. With such a huge amount of digital content, there will develop new businesses geared to organizing and protecting this content, creating new and personal entertainment, and allowing efficient sharing and even monetary exchange for the use of shared content. As personal content becomes larger than commercial content in the next decade, the economy of personal content will become the core business driver for much of the entertainment market. To drive these new businesses, the first step will be to develop sophisticated automatic metadata generation based upon digital content and also drawing upon the mass of accumulated personal content and experiences that are being created in our content archives. Condensing and using this personal database will take indexing and metadata creation into new directions ultimately making this content easier to use and more useful for the individual as well as enabling more efficient and useful sharing of this content with others. The same increase in organization and personal information could also allow new business models allowing individuals to use this personal database in new ways to entertain, educate, and do business. Ultimately we may all use these personal databases or life-logs to create virtual personal assistants that can remind us who all the people in a meeting are that we have met before as well as their backgrounds and relationships as we have experienced them. New standards for capturing and processing this personal data will be necessary, and this could be provided in the storage utility. Very large storage requirements will be needed. It is likely that by 2015 we could see homes with a petabyte of digital storage (1015 bytes) and most of this being personal content, copies of various personal and commercial content, and local backup of this content. Homes may also have terabytes (1012 bytes) of data remotely accessed over the internet for remote backup or continuous data protection, which provides disaster recovery, content sharing, remote access, and synchronization between devices. These services may be provided by devices sold at retail that conform to new standards enabling these functions or they may be part of a service package by a service provider such as a cable company or telco. In the first case there is likely a one-time charge to purchase the equipment and for remote access a monthly fee, in the latter case for a somewhat higher monthly fee all the hardware costs would be included.

www.newnespress.com

The Future of Home Digital Storage

249

Service providers have found that adding DVR and other special functionality increase the stickiness of their services (i.e., decrease the likelihood that a customer may switch to some other service). Likely service providers will find that if they provide a system that offers entertainment as well as the overall storage utility discussed here that they create the ultimate sticky business, since they would provide whole house data services for their customers.

9.8 Projections for Storage Demands in New Applications In this section we will look at some analyst projections of digital storage needs and projected growth of different digital storage devices for consumer applications. In addition we will make some estimates of total consumer content growth over the next few years based upon now existing consumer devices as well as projected devices such as life-logs and personal memory assistants. We will also factor in the growth of storage required to support content sharing developed earlier in this chapter. First it is important to understand the impact the age of a user has on the types of consumer devices that he or she uses. Survey results in the first quarter of 2007 from iSuppli gave an interesting breakdown in the types of portable media devices owned versus age. These are shown in Figure 9.4. The general trend here is that more modern 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

CD Player

Mobile Handset

65+

60–64

55–59

50–54

45–49

40–44

35–39

30–34

25–29

18–24

PMP/MP3 Player

Age

Figure 9.4: Q1 2007 Survey Results: Primary Portable Music Player as a Function of User Age.3 3

Courtesy of iSuppli 2007.

www.newnespress.com

250

Chapter 9

players such as PMPs, MP3 players, or cell phones with player capability are more likely with younger users. This may be due to familiarity with CDs or large CD collections by older users or general resistance to newer technology. In any event it is pretty clear that younger users tend to be the most likely to adopt newer consumer technology. Figure 9.5 gives the iSuppli projection for the growth of digital audio video (AV) PMPs and MP3 players, broken down by whether they use hard disk drives or flash memory. iSuppli projects strong growth in this type of device through 2009 and then expects the market to begin saturating in 2010 and 2011. The market for small form factor hard disk drive players is expected to grow with the market for PMPs, since hard disk drives provide more cost-effective very high capacity storage required for keeping very large video and high resolution music libraries or for keeping much higher resolution video content than is possible for flash. It is expected that this will continue to be true for many years to come as both flash memory and hard disk drives vie for some of the same markets. Figure 9.6 gives iSuppli’s projections for flash memory capacity points for these AV players. Also indicated on the chart is the average flash memory of shipping products at the time the chart was created. The chart shows that expected average capacity will grow from about 2.5 GB in 2007 to just above 25 GB in 2011 with flash-based AV players with over 130 GB capacity available at that time. Note that 1.8-inch hard disk drives with capacities of over 400 GB could be available at that time for hard disk drive PMP devices.

Millions of Units

300

HDD PMP

250 200

HDD Non-PMP

150

Flash PMP

100 50

Flash Non-PMP 2011

2010

2009

2008

2007

2006

2005

0

Figure 9.5: Projection for AV Player Growth by Type of Digital Mass Storage Used.4 4

Courtesy of iSuppli 2007.

The Future of Home Digital Storage 100%

251

30,000

90% 25,000 80% 70% 20,000

50%

15,000

40% 10,000 30% 20% 5,000

MB

60%

131,072 65,536 32,768 16,384 8,192 6,144 4,096 2,048 1,024 512 256 128 64 Avg MB/player

10% 0% 2005

2006

2007

2008

2009

2010

2011

Figure 9.6: AV Flash-Based Player Storage Capacity Breakdown Including Average Storage Capacity of Players.5

Figure 9.7 gives the history and projections for unit growth of hard disk drives for various consumer applications. While hard disk drives are still expected to play a role in mobile applications, about 75 percent of total expected consumer electronics hard disk drive unit volume in 2012 is expected to be from in-home applications such as game systems and digital video recorders as well as direct attached and network attached external storage systems. By 2010 over 20 percent of all hard disk drive storage shipped (in bytes) will be used in consumer applications. Total disk drive storage shipments in 2010 should be about 182 exabytes so 20 percent is about 36 exabytes. In 2010 total expected shipments of flash memory are expected to be about 16 exabytes with about 9.7 exabytes of this expected to be used in consumer applications (including cell phones). All together, by 2010, hard disk drive and flash memory storage capacity for all consumer markets will total about 46 exabytes.

5

Courtesy of iSuppli 2007.

252

Chapter 9

200,000

Cell Phones

180,000

Automotive Digital Still Camera

Units in Thousands

160,000

Digital Video Camera 140,000

AV Players

120,000

PVR/DVR/STB/Home Network Games

100,000

Other Consumer Electronics

80,000 60,000 40,000 20,000 0 2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Figure 9.7: Projections for Hard Disk Drives in Consumer Applications.6

Optical storage media shipments are slowing in unit growth as downloading of digital content becomes more common. Only the high capacity optical disks will continue to grow for the distribution of very rich content that would take too long to send electronically. Also, people who prefer to have a tangible copy of their content will be more inclined to use optical storage. As indicated in the earlier survey these are more likely to be the older segment of the population. Based upon market trends in past years, we can make a projection of total exabytes shipped for consumer electronics applications. This projection, breaking down shipments of flash, hard disk drive, and optical shipment for consumer applications through 2012, is shown in Figure 9.8. As can been seen in this figure, optical disc shipments represent the largest segment of digital capacity in this period (the pickup after 2008 is due to growth in shipments of blue laser optical discs), although hard disk drive and flash memory capacity constitute over 35 percent of the total shipping storage capacity by 2012.

6

Source: Coughlin Associates, 2007.

www.newnespress.com

The Future of Home Digital Storage

253

300

Optical Exabytes Shipped

250

Flash Hard Disk Drive

200 150 100 50 0 2006

2007

2008

2009

2010

2011

2012

Year

Figure 9.8: Projections of Exabytes Shipped for Consumer Electronic Applications Through 2012 Including Optical Discs, Hard Disk Drives, and Flash Memory.7 By 2012 and thereafter, the effects of social networking and the growing collection and sharing capabilities expected for personal content as well as growth in the use of lifelogs and other such memory intensive products will likely drive storage demand (especially for hard disks and flash memory) at a much higher rate than existed prior to 2012. Next, we make an estimate of the total storage capacity required for personal and commercial content in consumer environments (both static and mobile consumer electronics products). We shall also include a factor for digital content due to shared noncommercial content. The results of this estimate are given in Table 9.6. If we accept the projections in Table 9.6 and also Figure 9.8 then by 2011 utilization of all shipped capacity will be over 90 percent, and by 2012 the demand for storage will exceed shipping supply. Thus it is quite possible that demand for storage product total shipped capacity could be greater in the next decade than projected in Figure 9.8. If we assume 70 percent utilization of shipped storage with new content by 2015, then the actual digital storage capacity of shipped storage products required per Table 9.6 will be about 4,300 exabytes or 4.3 zetabytes. This table depends upon a number of demand factors that have not yet surfaced, so actual growth in digital storage demand may be more or less aggressive. However, the potential is present for very large consumer electronics demand growth in the coming years. 7

Source: Coughlin Associates, 2007.

www.newnespress.com

254

Chapter 9 Table 9.6: Estimated Growth of Personal and Commercial Content Stored or Associated with Consumer Static and Mobile Consumer Devices (Storage Units in Exabytes) Year

Commercial Content

Self Generated Personal Content

Shared Personal Content

Total

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

4 8 16 30 48 69 93 120 150 184

5 9 13 24 35 113 274 603 1,279 2,664

0 0 0 1 3 7 17 39 88 194

9 17 29 55 86 189 384 762 1,517 3,041

By 2015, mobile hard disk drives with over a terabyte of memory, including several hundred gigabytes of flash memory, should be available for use in portable products such as life-logs and personal memory assistants. Likewise, mass storage devices for the home will use hard disk drives with many terabytes of digital storage capacity. These drives could be used in integrated home network storage systems or the home storage utility outlined here. Optical discs could be available with terabyte storage capacity to bring very high resolution content into the home. By 2015, we might even see new solid state storage technologies used in some consumer products. Whatever else happens in the future, as long as there are people, there will be ever increasing demand to entertain ourselves and to store and utilize our life experiences. The human need to know more and share more will be an enormous driver for storage capacity as well as storage performance growth.

9.9 Digital Storage as Our Cultural Heritage There is now more information available in digital form than ever existed in earlier analog manifestations. Digital storage has enabled inexpensive capture, editing, and retention of human content in ways that have never existed before in the history of

www.newnespress.com

The Future of Home Digital Storage

255

mankind. In a very real sense digital storage preserves our current civilization and the human experience for future generations. Digital storage is the means by which we will pass on our cultural heritage; it is similar to but richer than the older methods such as books or pictures. However, this prediction coming true requires that this content actually gets passed down. There are real challenges in managing and backing up content, particularly personal content in the home, and making sure that this content makes it to later points in time in a format and a medium that can be read. Many people don’t throw older unused content away or eliminate multiple copies of content. This makes the importance of automated systems that can help people organize and protect their personal and other valuable content even more important. Without the development of such helpful robots as in the home storage utility, most people will drown in a sea of their own recorded experiences. Another important element in preserving this content for the long term is to have online copies that are available to recover and restore this content in case your home or your storage devices are damaged or destroyed. In order to make this option popular, encryption of the content to protect the privacy of personal content will be required, so that the content can still be used even if all of your electronic equipment is lost. This may require some biometric component in the security to allow the reestablishment of content access. The development of a personal memory assistant will also help protect the most valuable personal data by keeping a copy of at least proxy data with the user. Taking this concept a bit further, could a collection of personal digital histories from a point in time be used together to create some sort of representation of collective personal experiences of that time? Could such content make history more interesting to learn and lead to greater insights into large and small events in history? Can this be done in such a way as to protect more personal content or to make the individuals anonymous? Once collected and protected, digital histories will be part of the tools available to future generations to understand past times. If we can learn from history, perhaps the additional introspection and understanding that access to such content could make possible will lead to a greater sensitivity to the opinions and thoughts of others. Maybe this could help us be better people. It could certainly make us more responsible for our acts.

www.newnespress.com

256

Chapter 9

9.10 Chapter Summary •

Sharing content within the home and through the internet will create very large storage demand. Combined with the transition of individuals from just being consumers of content to be active generators of content, the required digital storage will be enormous.

•

Inexpensive digital storage, as well as faster and more complex microprocessors and electronics, will enable the creation of life-logs and consequently personal memory assistants. These combine inexpensive and pervasive personal experience storage with indexing and organization in the home storage utility to create a device that can help you recover past personal history and associated information when it is needed. Such a device will be useful for medical as well as business and personal purposes.

•

In order to simplify the management and use of digital content in the home, various paths to home storage virtualization will be developed. Over several phases, this will eventually lead to a unified home storage utility.

•

The home storage utility will provide efficient management of physical storage resources in and around the home including automobiles and mobile devices. The storage utility will organize digital content, make sure that it is backed up in an efficient and secure manner, and make sure that the content is synchronized with the various devices so that content is where it is needed. In the case of life-logs and other intensive content, the storage utility should create a “life script” and generate automatic metadata allowing efficient finding and use of the personal content.

•

The home storage utility and home networking in general, including home media centers, should provide proper protection of commercial and quasi-commercial content as well as allow sharing while still preserving privacy of personal content.

•

Large organized databases of personal content and experiences will lead to new business models for entertainment and education. They will also lead to new challenges for long term data retention.

•

These large personal databases and content caches will also have personal and economic value as well as collectively giving future researchers greater insight into the way we lived and the things we did.

www.newnespress.com

CHAPTE R 10

Standards for Consumer Electronics Storage Objectives in This Chapter •

Become familiar with standards for digital storage.

•

Learn about consumer product standards.

•

Be introduced to home networking standards.

•

Explore needed standards for future consumer electronics development.

10.1 Digital Storage Standards There are many standards bodies involved in digital storage related standards. Designers of digital storage for consumer devices benefit from being familiar with these standards or even joining these standards efforts. Following are some resources to learn more about various useful standards.

10.1.1 ANSI T13 Committee Technical Committee T13 is responsible for all interface standards relating to the popular AT Attachment (ATA) storage interface. This interface is utilized as the disk drive interface on most personal and mobile computers today including most consumer electronics devices. The charter of Technical Committee T13 calls for it to provide a public forum for the development and enhancement of storage interface standards for high volume personal

www.newnespress.com

258

Chapter 10

computers. The work of T13 is open to all materially impacted individuals and organizations.1

10.1.2 CE-ATA Standard Leading industry consumer and storage companies are spearheading the development of a storage interface tailored to the needs of the handheld and consumer electronics (CE) market segments. The promoter companies for the new initiative are Hitachi, Intel, Marvell, Seagate, and Toshiba. The objective of the initiative is to define a standard interface for small form factor disk drives that addresses the requirements of the handheld and consumer electronics market segments, including low pin count, low voltage, power efficiency, cost-effectiveness, and integration efficiency. The creation of the CE-ATA technology is necessary to address these specific requirements and respond to the needs of today’s quickly changing environment.2

10.1.3 Serial ATA (SATA) and eSATA Standards The Serial ATA International Organization (SATA-IO) is an independent, nonprofit organization developed by and for leading industry companies. Officially formed in July 2004 by incorporating the previous Serial ATA Working Group, the SATA-IO provides the industry with guidance and support for implementing the SATA specification. The standardized SATA specification replaces the 15-year-old parallel AT technology with a high speed serial bus supporting up to 10 years of expected future. Initially SATA was designed as an internal or inside-the-box interface technology, bringing improved performance and new features to internal PC or consumer storage. Creative designers quickly realized the innovative interface could reliably be expanded outside the PC, bringing the same performance and features to external storage needs instead of relying on USB or 1394 interfaces. Called external SATA or eSATA, customers can now utilize shielded cable lengths up to 2 m outside the PC to take advantage of the benefits the SATA interface brings to storage. SATA is now out of the box as an external standard, with specifically defined cables, connectors, and signal requirements released as new standards in mid-2004. eSATA provides more performance than existing solutions and is hot pluggable. 1 2

URL: http://www.t13.org. URL: http://www.ce-ata.org.

www.newnespress.com

Standards for Consumer Electronics Storage

259

•

Key benefits of eSATA: Up to six times faster than the existing external storage solutions, USB 2.0, and 1394.

•

Robust and user friendly external connection.

•

High performance, cost-effective expansion storage.

•

Up to 2 m shielded cables and connectors.3

10.1.4 iVDR Standard The iVDR Consortium is an international forum that has developed and promotes the adoption of a removable hard disk drive standard. iVDR stands for Information Versatile Disk for Removable use. The iVDR Consortium seeks to have the iVDR standards adopted to create a removable storage standard for computer and consumer electronics applications. 50 companies are currently members. Many of these companies are from Japan where the standard was originally developed. In 2007 the first products using iVDR were announced for the Japanese market. By the end of 2007 it is expected that the first products in the U.S. using this technology will be announced. Two specifications have been set for iVDR, one for a cartridge device designed for portability and the other for embedded storage that is mounted inside equipment. The cartridge specification allows for transfers of content between equipment in homes, on vehicles, or in portable equipment. The embedded specification governs an approach to adding hard disks into digital TVs or DVRs. Both types while being removable, enable large capacity recording and high speed random access which are possible only with hard disk drives. The physical and logical interface used in iVDR is based on Serial ATA and allows high speed transfer rate of 1.5 GBps or 3.0 GBps. By using multiple iVDR units, a storage server with several TB of removable storage can be created. Content security is dealt with using a dedicated technology called Security Architecture for Intelligent Attachment (SAFIA). iVDRs mounting this SAFIA are called iVDR-Secure. The organization hopes to have SAFIA be a means for iVDR to become a widely used standard for legitimate sharing and use of protected content by consumers.4 3 4

URL: http://www.sata-io.org. URL: http://www.ivdr.org.

www.newnespress.com

260

Chapter 10

10.1.5 Open NAND Flash (ONFI) Standard The Open NAND Flash Interface Working Group is a consortium of technology companies working to develop open standards for NAND flash memory chips and devices that communicate with them. The formation of ONFI was announced at the Intel Developer Forum in March 2006. The group’s goals notably do not include the development of a new consumer flash memory card format. Rather, ONFI seeks to standardize the low-level interface to raw NAND flash chips. The ONFI consortium is led by several prominent manufacturers of NAND flash memory: Hynix, Intel, Micron Technology, Phison, Sony, and STMicroelectronics. SanDisk and Samsung are not currently members. ONFI has produced a specification for a standard interface to NAND flash chips. Version 1.0 of this specification was released on December 28, 2006, and is available at no cost from the ONFI web site. It specifies: •

A standard physical interface (pin-out) for NAND flash in TSOP-48, WSOP-48, LGA-52, and BGA-63 packages.

•

A standard mechanism for NAND chips to identify themselves and describe their capabilities (comparable to the Serial Presence Detection feature of SDRAM chips).

•

A standard command set for reading, writing, and erasing NAND flash.

•

Standard timing requirements for NAND flash.

•

Improved performance via a standard implementation of read cache and increased concurrency for NAND flash operations.

•

Improved data integrity by allowing optional ECC memory features.5

10.1.6 USB Smart Drives The USB Flash Drive Alliance (UFDA), a consortium of leading USB flash drive manufacturers, supports USB smart drives, which allow users to run active programs from USB flash drives. USB smart drives provide users similar functionality to desktop applications run from a flash drive. 5

URL: http://www.onfi.org.

www.newnespress.com

Standards for Consumer Electronics Storage

261

UFDA members include: Lexar, PNY Technologies, Inc., Samsung Semiconductor Inc., Buffalo, Crucial Technology, Infineon Technologies, International Microsystems Inc., Kingston Technology, and Peripheral Enhancement. Another initiative to create smart USB devices is the U3 initiative from SanDisk. This approach has similar goals as the UFDA but it is championed by SanDisk. Recently SanDisk and Microsoft have announced an agreement in products using U3. On May 11, 2007, Microsoft and SanDisk announced an agreement that expands on SanDisk’s existing U3 Smart Technology. Under the terms of the agreement, Microsoft will develop a new software experience and SanDisk will develop new hardware capabilities, including the addition of TrustedFlash security technology. The new offering will be commercially available starting in the second half of 2008.6

10.1.7 Flash Card Standards There are many types of removable storage cards used in consumer devices. These usually use flash memory, although other types of memory have been used in these devices, such as 1-inch hard disk drives in the compact flash form factor. Table 10.1 gives information on some of the most used flash card standards.

10.1.8 Universal Flash Storage Micron Technology, Nokia, Samsung Electronics, Sony Ericsson, Spansion, STMicroelectronics, and Texas Instruments are working to create the Universal Flash Storage (UFS) specification, a standardized format for removable flash memory cards and embedded memory technology. The UFS proposals will be standardized by the JEDEC Solid State Technology Association. JEDEC is an open-standards organization in the semiconductor industry. JEDEC is known for its expertise in the standardization of component technologies and solutions. UFS will provide a revolutionary leap toward supporting very low access times required for memories, as well as enabling high-speed access to large multimedia files, while reducing power-consumption in consumer electronic (CE) devices. The target performance level is expected to be a significant advancement beyond that of the varied flash cards popular today. Today, users experience a three-minute access time for a 6

URL: http://www.usbflashdrive.org/, http://www.u3.com.

www.newnespress.com

262

Chapter 10 Table 10.1: Flash Card Standards7

Name

Acronym

Form Factor (mm)

DRM

PC Card (http://www.pcmcia.org) CompactFlash I (http://www.compactflash. org) CompactFlash II (http://www.compactflash. org) SmartMedia (http://ssfdc.or.jp/english/) Memory Stick (http://www.memorystick.org) Memory Stick Duo (http://www. memorystick.org) Memory Stick Micro M2 (http://www. memorystick.org) Multimedia Card (http://www.mmca.org) Reduced Size Multimedia Card (http://www. mmca.org) MMCmicro Card (http://www.mmca.org) Secure Digital Card (http://www.sdcard.org) miniSD Card (http://www.sdcard.org) microSD card (http://www.sdcard.org) xD-Picture Card (http://www.xd-picture.com)

PCMCIA CF-I

85.6 × 54 × 3.3 43 × 36 × 3.3

None None

CF-II

43 × 36 × 5.5

None

SM / SMC MS MSD

45 × 37 × 0.76 50.0 × 21.5 × 2.8 31.0 × 20.0 × 1.6

None MagicGate MagicGate

M2

15.0 × 12.5 × 1.2

MagicGate

MMC RS-MMC

32 × 24 × 1.5 16 × 24 × 1.5

None None

MMCmicro SD miniSD microSD xD

12 × 14 × 1.1 32 × 24 × 2.1 21.5 × 20 × 1.4 11 × 15 × 1 20 × 25 × 1.7

None CPRM CPRM CPRM None

90-minute (4 gigabyte) high-definition movie; with the new standard, this would be reduced to a few seconds. Major applications such as mobile handsets, digital still cameras and other consumer electronic devices will benefit from the convenience of a universal open standard based specification that is intended to reduce the time-consuming process of enabling interoperability among the various types of removable and embedded memory solutions at the system level. Universal Flash Storage is planned to provide consumers in the future the convenience of a unified removable memory card that can be shared among various mobile, portable, and other CE devices without the need for any adapters. The Universal Flash Storage standard is expected to be finalized in 2009.8

10.1.9 Trusted Computing Group Standards The Trusted Computing Group (TCG) is a nonprofit industry organization formed to develop, define, and promote open standards for hardware-enabled trusted computing 7 8

Source: http://en.wikipedia.org/wiki/Memory_card. URL: www.jedec.org.

www.newnespress.com

Standards for Consumer Electronics Storage

263

and security technologies. Up until 2007, TCG standards deployed in products are the PC-based Trusted Platform Module (TPM) and the Trusted Software Stack (TSS). These two standards are designed into integrated circuits, systems, and applications. TPM/TSS are available on new PCs and laptops by being built into Intel and AMD microprocessors. These computers are manufactured by Acer, Dell, Fujitsu, HewlettPackard, and IBM. The TPM chips can be used for device authentication, rogue software detection, and secure storage. TCG has expanded upon the existing TPM model to create a set of Storage Work Group specifications. A key element in providing this security is encryption of data on the storage device with the encryption key kept in a non-user-accessible area on the storage device. There are three main security benefits of these specifications. 1.

It creates trust relationships between storage devices and hosts using mutual identity and authentication of hosts and storage devices. By extending TPM into storage devices, the device can limit who can read or write to the storage device.

2.

TCG-enabled storage enables secure control over storage device features. These products provide protected storage for specific users, systems, and applications. They also provide exclusive control of encrypted data on the TCG-enabled storage device.

3.

They enable secure communications between storage devices and their hosts using session-oriented security commands defined in security extensions of storage device command specifications such as SCSI and ATA.

Since data resides mostly on storage devices, these are a natural target for security. By encrypting the data and storing the key on the hard disk drive it becomes impossible to access this data through probe points and other weak points on encryption systems where the key is accessed off the disk drive. Also, by incorporating encryption on the drive, no additional hardware or software is required with the attendant risk that the hardware or software could fail. TPM is enabled by a tiny processor whose firmware cannot be modified. Security services are called through specific APIs that protect the storage device behind a trust boundary. Only trusted entities with the proper authorization to the API can use TCG trusted storage. The protection of encrypted content on a storage device can be used to provide secure backup of content between different TCG-enabled storage devices. The TCG specification is also able to perform pervasive logging, meaning that it can support logging and clocking capabilities. This can help in determining when data should be moved from one device to another based upon use.

www.newnespress.com

264

Chapter 10

Although the initial market for TCG storage devices is hard disk drives used in laptops, there is every reason that this technology can be used for other applications (including consumer electronics applications) and on other storage devices. TCG storage could be an important element in protecting personal content privacy or in DRM protection of personal or commercial content.9

10.2 Consumer Product Standards There are several important standards covering the operation of consumer products (ultimately some of these could incorporate the standards functions referred to in this book). There are also a great many options being developed for networking in the home that eventually will be used for sharing content between storage devices. We devote a section of this chapter to home network standards.

10.2.1 UHAPI NXP Semiconductors (formerly the semiconductor unit of Royal Philips Electronics) and the semiconductor unit of Samsung Electronics Co., Ltd. partnered to solve the task-level complexities that stall innovation and progress at all layers of the home consumer electronics stack. As a result, the Universal Home API (UHAPI) Forum was established to further this vision. The UHAPI Forum’s mission is to develop and maintain the A/V API for home consumer electronics products such as DVD players and recorders, televisions, media servers, media adapters, and set-top boxes. As industry-wide adoption is realized, the hardware-independent UHAPI will enable independent software vendors, system integrators, and consumer electronics product manufacturers to create middleware and application software that is easily ported across multiple UHAPI-compliant solutions. Latest specification is UHAPI 1.2.10

10.2.2 DLNA The Digital Living Network Alliance (DLNA) is focused on delivering an interoperability framework of design guidelines based on open industry standards to 9 10

URL: www.trustedcomputinggroup.com. URL: http://www.uhapi.org/.

www.newnespress.com

Standards for Consumer Electronics Storage

265

complete the cross-industry digital convergence. The DLNA Networked Device Interoperability Guidelines, expanded in October 2006, consists of three volumes covering Architecture & Protocols, Media Format Profiles, and Link Protection. It provides vendors with the information needed to build interoperable networked platforms and devices for the digital home. The digital home vision integrates the internet, mobile, and broadcast networks through a seamless, interoperable network. The latest document is DLNA Networked Device Interoperability Guidelines, October 2006.11

10.2.3 OSGi Alliance The Open Service Gateway Initiative (OSGi) Alliance is a worldwide consortium that advances a proven and mature process to assure interoperability of applications and services based on its component integration platform. The alliance provides specifications, reference implementations, test suites, and certification to foster a valuable cross-industry ecosystem. Adoption of the component-based platform reduces time-to-market and development costs because it enables integration of pre-built and pre-tested modules. It reduces maintenance costs and provides aftermarket opportunities because networks are used to dynamically update or deliver services and applications in the field. Alliance members represent diverse markets including SmartHome, automotive electronics, mobile, and enterprise. Member company industries include leading service and content providers, infrastructure/network operators, utilities, software developers, gateway suppliers, consumer electronics/device suppliers (wired and wireless), and research institutions. Latest OSGi Service Release 4.1 in May 2007.12

10.2.4 DRM Standards MagicGate is a copy-protection technology introduced by Sony in 1999 as part of the Secure Digital Music Initiative (SDMI). It works by encrypting the content on the

11 12

URL: http://www.dlna.org/en/industry/home/. URL: http://www2.osgi.org.

www.newnespress.com

266

Chapter 10

device and using MagicGate chips in both the storage device and the reader to enforce control over how files are copied. MagicGate encryption is used in the memory cards of the PlayStation 2 and, as of 2004, has been introduced into all of Sony’s Memory Stick products. Some devices, such as Sony’s Network Walkman, will only accept Memory Sticks that use MagicGate technology. Note that MagicGate is not supported by many consumer electronics companies and even many Sony products do not recognize cards with the MagicGate chips.13 Content Protection for Recordable Media and Pre-Recorded Media (CPRM/CPPM) is a mechanism for controlling the copying, moving, and deletion of digital media on a host device, such as a personal computer, or other digital player. It was developed and controlled by the 4C Entity, LLC (which consists of IBM, Intel, Matsushita, and Toshiba). The CPRM/CPPM specification is designed to balance the robustness and renewability requirements of content owners with ease of implementation for consumer electronics designers. The system defined by the specification relies on key management for interchangeable media, content encryption, and media-based renewability. The CPRM/CPPM specification defines a cryptographic method for protecting entertainment content recorded on physical media. The currently implemented method utilizes symmetric encryption with the Cryptomeria cipher (C2) algorithm. The types of physical media supported include recordable DVD media and flash memory. The most widespread use of CPRM is probably in Secure Digital cards.14

10.3 Home Networking Standards Following is a listing of various home wired and wireless networking standards with URLs where the reader can get more information.

10.3.1 Bluetooth Bluetooth wireless technology is a short-range communications technology intended to replace the cables connecting portable and/or fixed devices while maintaining high 13 14

URL: www.sony.net/Products/SC-HP/cx_news/vol20/pdf/tw.pdf. URL: www.4centity.com.

www.newnespress.com

Standards for Consumer Electronics Storage

267

levels of security. The key features of Bluetooth technology are robustness, low power, and low cost. The Bluetooth specification defines a uniform structure for a wide range of devices to connect and communicate with each other. Latest core specification is v2.1+EDR.15

10.3.2 CableHome CableHome is developing the interface specifications necessary to extend high-quality cable-based services to network devices within the home. Working with technology suppliers, the CableHome team addresses issues such as device interoperability, Quality of Service (QoS), and network management. CableHome 1.1 specification was released in July 2006.16

10.3.3 DOCSIS DOCSIS, the CableLabs® CertifiedTM Cable Modem project, also known as the Data Over Cable Service Interface Specification, defines interface requirements for cable modems involved in high-speed data distribution over cable television system networks. The certified cable modem project also provides cable modem equipment suppliers with a fast, market-oriented method for attaining cable industry acknowledgment of DOCSIS compliance and has resulted in high-speed modems being certified for retail sale. Latest interface specification is DOCIS 3.0.17

10.3.4 IEEE 1394 IEEE 1394 multimedia connection enables simple, low-cost, high-bandwidth isochronous (real-time) data interfacing between computers, peripherals, and consumer electronics products such as camcorders, VCRs, printers, PCs, TVs, and digital cameras. With IEEE 1394-compatible products and systems, users can transfer video or still images from a camera or camcorder to a printer, PC, or television, with no image degradation.18 15 16 17 18

URL: http://www.bluetooth.com. URL: http://www.cablelabs.com/projects/cablehome/. URL: http://www.cablemodem.com/. URL: http://www.1394ta.org/Technology/index.htm.

www.newnespress.com

268

Chapter 10

10.3.5 Wireless HD Wireless HD (WiHD) is a special interest group formed to develop a specification for a wireless high-definition digital interface that is intended to enable high-definition audio video (A/V) streaming and high-speed content transmission for consumer electronics devices. This is an in-room solution running at 60 GHz. It uses a steerable antenna array to form a beam that can transmit content up to 500 MBps within a 10 m distance. This more or less line-of-sight technology is intended to allow devices such as DVD players or DVRs to move HD content to a monitor without the use of HDMI wiring. Founding members of the group are LG, Panasonic, NEC Samsung, SiBeam, Sony, and Toshiba. Note that the IEEE has its own set of specifications that it is working on that allows even higher data rates using 90 GHz to support Ultra HD.19

10.3.6 IEEE 802 The IEEE 802 standard comprises a family of networking standards that cover the physical layer specifications of technologies from Ethernet to wireless. IEEE 802 is subdivided into 22 parts that cover the physical and data-link aspects of networking. The better known specifications include 802.3 Ethernet, 802.11 Wi-Fi, 802.15 Bluetooth/ZigBee, and 802.16. Table 10.2 lists highlights of the most popular sections of IEEE 802 as of the time of this writing. Copies of these specifications are available from the IEEE. The most popular specifications are in bold.20

10.3.7 OpenCable The OpenCable initiative, managed by the Advanced Platforms and Services group at CableLabs, began in 1997 with the goal of helping the cable industry deploy interactive services over cable. OpenCable specifications help achieve the goal of interactive services by meeting three key objectives: 1.

Define the next-generation digital consumer device.

2.

Encourage supplier competition.

3.

Create a retail hardware and software platform.

19 20

URL: http://www.wirelesshd.org. URL: www.ieee802.org.

www.newnespress.com

Standards for Consumer Electronics Storage

269

Table 10.2: IEEE 802 Networking Standards Standard

Topic of the Standard

Description

802 802.1

Overview Bridging

802.2

Logical Link

802.3

Ethernet

802.5

Token Ring

802.11

Wi-Fi

802.11a

—

802.11b

—

802.11d

—

Basics of Physical and Logical Networking Concepts LAN/MAN Bridging and Management. Covers Management and the Lower Sub-Layers of OSI Layer 2, Including MAC-Based Bridging (Media Access Control), Virtual LANs and Port-Based Access Control. Commonly Referred to as the LLC or Logical Link Control Specification. The LLC is the Top Sub-Layer in the DataLink Layer, OSI Layer 2. Interfaces with the Network Layer 3. “Grandaddy” of the 802 Specifications. Provides Asynchronous Networking Using “Carrier Sense, Multiple Access with Collision Detect” (CSMA/CD) over Coaxial Cable, Twisted-Pair Copper, and Fiber Media. Current Speeds Range from 10 MBps to 10 GBps. The Original Token-Passing Standard for Twisted-Pair, Shielded Copper Cables. Supports Copper and Fiber Cabling from 4 MBps to 100 MBps. Often Called “IBM Token-Ring.” Wireless LAN Media Access Control and Physical Layer Specification. 802.11a, b, g, etc. Are Amendments to the Original 802.11 Standard. Products That Implement 802.11 Standards Must Pass Tests and Are Referred to as “Wi-Fi Certified.” • Specifies a PHY That Operates in the 5 GHz U-NII Band in the U.S.—Initially 5.15–5.35 and 5.725–5.85— Since Expanded to Additional Frequencies • Uses Orthogonal Frequency-Division Multiplexing • Enhanced Data Speed to 54 MBps • Ratified After 802.11b • Enhancement to 802.11 That Added Higher Data Rate Modes to the DSSS (Direct Sequence Spread Spectrum) Already Defined in the Original 802.11 Standard • Boosted Data Speed to 11 MBps • 22 MHz Bandwidth Yields Three Non-Overlaping Channels in the Frequency Range of 2.400 GHz to 2.4835 GHz • Beacons at 1 MBps, Falls Back to 5.5, 2, or 1 MBps from 11 MBps Maximum • Enhancement to 802.11a and 802.11b That Allows for Global Roaming • Particulars Can Be Set at Media Access Control (MAC) Layer

www.newnespress.com

270

Chapter 10 Table 10.2: Continued

Standard

Topic of the Standard

Description

802.11e

—

802.11g

—

802.11h

—

802.11i

—

802.11j

—

802.11k

—

802.11m

—

802.11n

—

802.11x

—

802.12

Demand Priority

802.13 802.14

Not Used Cable Modems

802.15

Wireless Personal Area Networks

• Enhancement to 802.11 That Includes Quality of Service (QoS) Features • Facilitates Prioritization of Data, Voice, and Video Transmissions • Extends the Maximum Data Rate of WLAN Devices That Operate in the 2.4 GHz Band, in a Fashion That Permits Interoperation With 802.11b Devices • Uses OFDM Modulation (Orthogonal FDM) • Operates at up to 54 Megabits per Second (MBps), With Fall-Back Speeds That Include the “B” Speeds • Enhancement to 802.11a That Resolves Interference Issues • Dynamic Frequency Selection (DFS) • Transmit Power Control (TPC) • Enhancement to 802.11 That Offers Additional Security for WLAN Applications • Defines More Robust Encryption, Authentication, and Key Exchange, as Well as Options for Key Caching and Pre-Authentication • Japanese Regulatory Extensions to 802.11a Specification • Frequency Range 4.9 GHz to 5.0 GHz • Radio Resource Measurements for Networks Using 802.11 Family Specifications • Maintenance of 802.11 Family Specifications • Corrections and Amendments to Existing Documentation • Higher-Speed Standards—Under Development • Several Competing and Non-Compatible Technologies; Often Called “Pre-N” • Top Speeds Claimed of 108, 240, and 350+ MHz • Competing Proposals Come From the Groups EWC, Tgn Sync, and Wwise and Are All Variations Based on MIMO (Multiple Input, Multiple Output) • Misused “Generic” Term for 802.11 Family Specifications Increases Ethernet Data Rate to 100 MBps by Controlling Media Utilization. Not Used Withdrawn PAR. Standards Project No Longer Endorsed by the IEEE. Communications Specification That Was Approved in Early 2002 by the IEEE for Wireless Personal Area Networks (WPANs).

Standards for Consumer Electronics Storage

271

Table 10.2: Continued21 Standard

Topic of the Standard

Description

802.15.1

Bluetooth

802.15.3a 802.15.4 802.15.5

UWB ZifBee Mesh Network

802.16

Wireless Metropolitan Area Networks

802.17

Resilient Packet Ring Radio Regulatory TAG Coexistence Mobile Broadband Wireless Access Media Independent Handoff Wireless Regional Area Network

Short Range (10 m) Wireless Technology for Cordless Mouse, Keyboard, and Hands-Free Headset at 2.4 GHz Short Range, High-Bandwidth Ultra Wideband” Link Short Range Wireless Sensor Networks • Extension of Network Coverage Without Increasing the Transmit Power or the Receiver Sensitivity • Enhanced Reliability Via Route Redundancy • Easier Network Configuration—Better Device Battery Life This Family of Standards Covers Fixed and Mobile Broadband Wireless Access Methods Used to Create Wireless Metropolitan Area Networks (WMANs). Connects Base Stations to the Internet Using OFDM in Unlicensed (900 MHz, 2.4, 5.8 GHz) or licensed (700 MHz, 2.5–3.6 GHz) Frequency Bands. Products That Implement 802.16 Standards Can Undergo WiMAX Certification Testing. IEEE Working Group Description

802.18 802.19 802.20 802.21 802.22

IEEE 802.18 Standards Committee IEEE 802.19 Coexistence Technical Advisory Group IEEE 802.20 Mission and Project Scope IEEE 802.21 Mission and Project Scope IEEE 802.22 Mission and Project Scope

The OpenCable project has two key components: a hardware specification and a software specification. The hardware specification describes both one-way and two-way digital cable-ready “host” devices that are interoperable with cable systems throughout the U.S. The software specification of the OpenCable project, called the OpenCable Application Platform (OCAPTM), solves the problem of proprietary operating system software, thereby creating a common platform for interactive television applications and services. Interactive (bi-directional) OpenCable products require an OCAP middleware stack, but the OpenCable Unidirectional Receiver (OCUR) does not.22 21 22

Source: http://searchmobilecomputing.techtarget.com/sDefinition/0,290660,sid40_gci992311,00.html. URL: http://www.opencable.com.

272

Chapter 10

10.3.8 PacketCable PacketCable is a CableLabs-led initiative that is aimed at developing interoperable interface specifications for delivering advanced, real-time multimedia services over a two-way cable plant. PacketCable networks use Internet Protocol (IP) technology to enable a wide range of multimedia services, such as IP telephony, multimedia conferencing, interactive gaming, and general multimedia applications. PacketCable 2.0 is the most recent release.23

10.3.9 Voice over IP Voice over Internet Protocol is also called VoIP. Other names for this technology are IP Telephony, internet telephony, broadband telephony, and voice over broadband. This technology involves the routing of voice conversations over the internet or through other IP-based networks.24

10.3.10 Universal Plug and Play The Universal Plug and Play (UPnP) architecture offers pervasive peer-to-peer network connectivity of PCs of all form factors, intelligent appliances, and wireless devices. The UPnP architecture is a distributed, open networking architecture that leverages TCP/IP and the web to enable seamless proximity networking in addition to control and data transfer among networked devices in the home, office, and everywhere in between. UPnP provides the following capabilities:

23 24

•

Media and Device Independence: UPnP technology can run on any network technology including Wi-Fi, coaxial cable, phone line, power line, Ethernet, and 1394.

•

Platform Independence: Vendors can use any operating system and any programming language to build UPnP products.

•

Internet-Based Technologies: UPnP technology is built upon IP, TCP, UDP, HTTP, and XML, among others.

URL: http://www.packetcable.com. URL: http://www.tiaonline.org/standards/technology/voip.

www.newnespress.com

Standards for Consumer Electronics Storage

273

•

UI Control: UPnP architecture enables vendor control over device user interface and interaction using the browser.

•

Programmatic Control: UPnP architecture enables conventional application programmatic control.

•

Common: Vendors agree on base protocol sets on a per-device basis.

•

Extendable: Each UPnP product can have value-added services layered on top of the basic device architecture by the individual manufacturers.25

10.4 Needed Standards for Future Consumer Electronics Development Following are some ideas for standards that will be needed to carry out some of the initiatives described in this book. These range from creating applications integrated into storage devices to creating the integrated home storage utility.

10.4.1 Proposal for Open Standards for Storage Integration into Consumer Electronics The author believes that the best approach to facilitate the development of integrated applications in the digital storage device is to create a consortium to develop open integration standards that allow both the consumer electronics companies as well as the storage device manufacturers to maintain the means of creating proprietary products while improving the cost and performance of these products and increasing the ease and speed with which consumer end products can be created. A useful standard should also effectively deal with the issues of control of the end product and its functionality to the mutual satisfaction of the consumer electronics company and the digital storage device company. Development and implementation of such a standard could be beneficial to the host company by allowing rapid product design with better performance and lower overall cost. It would be beneficial to the storage device companies by allowing them to increase the value added quality of their products and thus improve profit margins. Following is a proposal and some expectations for the creation of such an open standard for consumer application integration into storage devices: 25

URL: www.upnp.org.

www.newnespress.com

274

Chapter 10

1.

The basic idea is to develop an open standard for the design, testing, and manufacturing of hard disk drives (and eventually other storage products) with integrated electronic application modules contained on the storage device electronics. These devices can be controlled by firmware and appropriate interfaces with the outside environment to create customizable proprietary consumer electronics applications.

2.

The standard must allow for the updating and importation of firmware for controlling the resulting applications or for reconfiguring between applications when appropriate.

3.

The standard must create flexible and yet defined rules for the incorporation of standard consumer application interfaces on the circuit board of the storage device to connect with the application housing.

4.

This approach should save considerably on the bill of materials costs of the resulting consumer electronics devices.

5.

This approach should also save considerably on the manufacturing costs of the resulting consumer electronics devices through use of a single manufacturing line and an integrated drive and application test.

6.

In addition, tight integration of the storage device with the host application should allow considerable improvement in host device performance as well as improvement in such factors such as power usage.

7.

This standard could also include other ancillary capabilities that could be incorporated into the integrated consumer applications such as the application SMART that can log application performance in a storage device log file and flag potential failures of the application modules.

10.4.2 Standards for Personal Content Metadata and Organization of Personal Content As our personal content grows, it will quickly become unmanageable unless companies develop ways to automatically organize and manage that data. For this reason we need to develop metadata format standards and processes for creating that metadata that employ automatic methods. These are needed because for personal content most users do not have the time to organize the digital files that result from living their lives. These standards should be incorporated as much as possible with the

www.newnespress.com

Standards for Consumer Electronics Storage

275

applications that use the content in order to create a process that works seamlessly with the applications. Likewise, we need search and indexing capabilities increasingly built either into applications or in a location, like an integrated home network, where this organization and management takes place. It could be that applications in static or mobile devices do the initial metadata capture and organization with deep indexing and associations created on a centralized and more powerful computer like that in a centralized home server appliance. Once organized, the metadata and proxies of the content can be loaded back on the mobile devices so that they can act as a personal memory assistant. These functions would also benefit from established metadata standards. In addition to standards, if personal content, particularly with life-logs, increase at the rate projected in this book, then we will need legal safeguards for this personal content. Privacy safeguards as well as protection of this content need to be established practices. Perhaps privacy can be dealt with using encryption technologies. In addition, should this personal content be open to court orders, or could a user witholding such content plead the Fifth Amendment?

10.4.3 Standards for Virtualization of Consumer Storage and the Creation of a Home Storage Utility At the enterprise storage level, the Storage Networking Industry Association (SNIA) has begun steps toward standardization of storage management. SNIA’s Storage Management Initiative Specification (SMI-S) has been widely adopted as a storage resource management standard, and laid the foundation for management software to be able to be written to one set of interfaces defined by an SMI-S profile for each type of device. Although this effort is in its early stages and enterprise storage has significant differences from home network storage, there are still significant amounts of overlap in the general functions. As a consequence, as the complexity of devices with storage as well as the sheer volume of digital content in the home increases, automated management processes will be needed. It is likely that some of the standards and practices developed initially to support enterprise network storage will eventually migrate to support networked devices with storage in the home. Virtualization of enterprise computer and storage assets is becoming a significant means to use these assets more effectively. Disk storage utilization in the enterprise is often at

www.newnespress.com

276

Chapter 10

50 percent or less. Using virtualization of available storage in storage arrays, a common managed pool combined with technologies such as deduplication may make it possible to increase overall disk utilization to greater than 70 percent. This lowers both the purchase price of the storage hardware and the cost of maintaining the data center utility (e.g., fewer storage arrays means less heat generated that air conditioning has to remove). Some sort of virtualization of the storage in all the devices in and around the home with storage and connected to the network could allow more effective management of both the storage resources as well as personal and commercial content. This unified management would help in efficient backing up of content, moving content between storage devices, and making the most effective use of the network bandwidth to provide content where it is needed. The storage networking standards developed by organizations such as SNIA combined with special capabilities for the unique features of the home storage environment will help bring about a unified home storage pool or utility. Development of the metadata standards and automated generation of metadata will also help in content management. These standards will help the system determine where particular content should go and who should have access to it. Also these capabilities could enable entertainment using a home network that includes personal as well as commercial content.

www.newnespress.com

APPE NDIX A

Table A.1: Specification Comparison of Some 1.8-Inch Hard Disk Drives and a Flash Solid State Drive Hitachi C4K60-30 Consumer Electronics Version

Samsung 1.8-inch PATA SSD

Units

1.2 1.4 1.4 1.4 0.6 0.2 0.08

1.4 1.1 1 1.1 0.25 0.08 0.07

0.4 0.4 0.4 0.4 0.1 0.01 0.01

W W W W W W W

5 to 60 −20 to 65 −40 to 70 15

5 to 60 −40 to 70 — 20

−25 to 85 −40 to 85 — N/A

°C °C °C °C/hour

8 to 90 8 to 90

5 to 90 5 to 95

N/A N/A

% %

29 40

29 40

N/A N/A

°C °C

Toshiba MK6006GAH

Power at 3.3 V Start (Max) Seek (Ave) Read (Ave) Write (Ave) Idle (Ave) Standby (Ave) Sleep (Ave) Ambient Temperature Operating Non-operating Storage Temperature Gradient (Max) Relative Humidity Operational Non-operating Max Wet Bulb Operational Non-operating

www.newnespress.com

278

Appendix A Table A.1: Continued Toshiba MK6006GAH

Hitachi C4K60-30 Consumer Electronics Version

Samsung 1.8-inch PATA SSD

Units

16 G

—

25.40 mm p-p max (5–10 Hz) 5.0 G max (10–500 Hz)

1.0 mm p-p max (5–22 Hz) 1.0 G max (22–500 Hz) 5.0 mm p-p max (5–22 Hz) 5.0 G max (22–500 Hz)

16 G

—

Shock Operational

500 (2.0 ms)

500 (2.0 ms)

G

Non-operating

1,500 (1.0 ms)

1,200 (1.0 ms)

1,000 (2.0 ms) 2,000 (1.0 ms)

— —

1.6 2.2

0 0

bels bels

−300 to 3,000 −400 to 15,000 — 1.00 × 1013

3,000 12,000 300 1.00 × 10−12

N/A N/A N/A —

m m m/min —

—

1,500

—

µTesla (DC)

300,000 N/A

— N/A

1,000,000 100,000 (Full Array)

Hours Cycles

Vibration Operational

Non-operating

Acoustic Noise Idle Seek Altitude Operational Non-operating Gradient Uncorrectable Error Rate External Magnetic Field MTBF R/W Cycles

1.0 G (5–500 Hz)

www.newnespress.com

G

APPE NDIX B

Home Networking Technology Trade Groups The HomePlug Alliance was founded to support data transport over power lines. Alliance members include Panasonic, Sharp Corp., Conexant Systems Inc., and Intellon Corporation. Embedded HomePlug 1.0 chipsets add about $10 to the cost of a device, and $20 to $30 to the retail price. These prices are expected to fall one to two years after introduction once volumes increase. The sustained data rate reported for the HomePlug 1.0 specification is 14 MBps (∼1.75 MBps). The HomePlug AV specification will allow for 200 MBps data rates or about 25 MBps.1 The HomePNA standard was developed by the Home Phoneline Networking Alliance. It was founded in June 1998, and is led by seven companies (Agere Systems, AT&T, Broadcom, Conexant, Hewlett-Packard Co., Motorola, and 2Wire). HomePNA 2.0 was developed to run over existing phone line and coaxial wiring in a symmetrical mode at a peak data rate up to 32 Mbps, with an average throughput of 10 MBps (∼1.25 MBps). The HPNA 2.0 network uses a shared physical media (wiring) and has no need for a switch or hub.2 The Multimedia over Coax Alliance or MoCA is based upon the Entropic out-ofband IP transport over coaxial cable technology. A trade group has been formed around this standard with membership including Linksys, Comcast, Echostar, Entropic, Panasonic, Motorola, Radio Shack, and Toshiba.3

1 2 3

URL: www.homeplug.org. URL: www.homepna.org. URL: www.mocalliance.org.

www.newnespress.com

280

Appendix B

The High Definition Audio-Video Network Alliance or HANA uses the IEEE 1394 standard over coaxial cable to create home networking. The user interface (UI) is delivered over the 1394 interface using well-established Web technologies—browsers and Web servers. A connected device is selected using the DTV’s own remote control, whereupon that device serves up its own UI for display on the DTV. Devices do not need to have custom software drivers installed to interoperate. If they are HANA devices, they are guaranteed to interoperate.4

4

URL: www.hanaalliance.org.

www.newnespress.com

APPE NDIX C

Companies Making Various Storage Products Used in Consumer Applications Flash Memory Manufacturers There are many manufacturers of flash memory used in consumer applications. These companies are located throughout the world, although the largest volume of flash memory is produced in Asia. Following is a list of many of the major suppliers of flash memory. •

Atmel offers parallel flash memories in a variety of configurations and architectures. Atmel flash memory is used in PC peripherals, mobile phones, networking equipment, and set-top boxes. (URL: http://www.atmel.com/ products/Flash/.)

•

Intel makes both NOR and NAND flash devices with single level and multilevel cells. Intel is combining much of its flash memory technology in a joint venture with STMicroelectronics. (URL: www.intel.com/design/flash/index.htm.)

•

Micron uses floating gate NAND cells and a standard interface, enabling pin and function drop-in compatibility and easy integration into most existing designs. Micron flash memory also offers high-performance features, including a READ CACHE function that streams more than 30 MBps. (URL: www.micron.com/ products/nand/.)

•

Samsung is the largest manufacturer of NAND flash memory in the world. Samsung flash is used in many mobile consumer products such as Apple iPods and iPhones. (URL: www.samsung.com/global/business/semiconductor/products/ flash/Products_Flash.html.)

www.newnespress.com

282

Appendix C

•

SanDisk was an early pioneer in flash memory and the inventor of flash memory cards. SanDisk has the biggest retail presence in the card memory market, especially in the U.S. SanDisk manufactures its flash memory in factories jointly owned with Toshiba. (URL: www.sandisk.com.)

•

Sharp offers a line of NOR flash memory products used in many applications for Execute In Place (XIP) applications such as in cell phones. (URL: www. sharpsma.com/Page.aspx/americas/en/726c69b4-8d0b-4cff-ac86-2b607e67c8ff/ Memory/.)

•

Spansion is a joint venture between AMD and Fujitsu. It is a major supplier of NOR flash memory used in many consumer applications. (URL: www.spansion.com.)

•

STMicroelectronics offers a complete range of NAND flash memory products. These are used in a number of consumer applications. STMicroelectronics NAND Flash products are offered with a 1.8 V and 3 V power supply and a number of packaging choices. (URL: www.st.com/stonline/products/families/ memories/fl_nand/index.htm.)

•

Toshiba invented flash memory and has remained in this market ever since. They produce a great variety of flash memory including NAND flash cards. They jointly own flash production facilities with SanDisk. (URL: www.toshiba.com/taec/.)

Hard Disk Manufacturers The hard disk drive industry has undergone significant consolidation over the years. Where there once were close to 50 companies making hard disk drives, there are now only 6. The listing of products and markets in the descriptions that follow are current as of the time of writing and will probably change with time. Likewise, there may not be as many hard disk drive companies in the future as there are now. •

Fujitsu makes 2.5-inch hard disk drives used in enterprise applications as well as mobile computers. They have had a limited exposure to the consumer electronics market. (URL: us.fujitsu.com/fcpa.)

•

Hitachi GST makes hard drives of all form factors from 1.0-inch to 3.5-inch drives. They pioneered the 1-inch drive in 1997. Hitachi provides disk drives to the static and mobile consumer electronics market. (URL: www.hitachigst.com.)

www.newnespress.com

Appendix C

283

•

Toshiba makes small form factor disk drives from 0.85-inch to 2.5-inch. They developed the 0.85-inch drive for the mobile phone market. Toshiba is the largest supplier of 1.8-inch disk drives for PMPs, and thus a major supplier of hard disk drives for consumer electronics applications. (URL: sdd.toshiba.com.)

•

Samsung makes 3.5-inch as well as 2.5-inch hard disk drives. Most of these are used in desktop and laptop computers, but they also provide drives for consumer electronics applications. (URL: www.samsung.com/global/business/hdd/.)

•

Seagate is the largest disk drive company. They make the full range of hard disk drive form factors. They also are the biggest supplier of hard disk drives for the consumer electronics industry. (URL: www.seagate.com.)

•

Western Digital is the second largest hard disk drive company, making 3.5-inch and 2.5-inch hard disk drives used in computer and consumer applications. They have been a major supplier of hard disk drives for game machines. (URL: www.westerndigital.com.)

Optical Disc Manufacturers Optical disk drives and media are manufactured by many companies. The actual manufacturing is generally in Asia, particularly in Taiwan or China. Most of the major consumer electronics companies also manufacture or at least have their name on optical disc drives. Some of the companies that make optical disc drives include: Panasonic, Samsung, Sharp, Sony, and Toshiba. HP introduced the Lightscribe technology that allows computers to use the laser in the optical drive writer to write a label on the side of the disc opposite to the data surface. External Storage Manufacturers There are a great many companies making various external storage products with USB, IEEE 1394, or eSATA interfaces. Most of these use hard disk drives either singly or in small arrays. Several of the hard disk drive manufacturers do this such as Hitachi, Seagate, Toshiba, and Western Digital. Many other companies buy hard drives and add them to external enclosure boxes or make and sell these boxes to consumers. Among the external storage suppliers without an in-company source of disk drives are Buffalo, HP, Iomega, Lacie, Data Robotics, and many others.

www.newnespress.com

This page intentionally left blank

Bibliography A Note on Sources What follows are various sources used in the process of creating this book. Some sources were used throughout the book and hence they are listed under general sources. Others were useful for particular chapters, and they are listed by chapter.

General S. Kipp. Broadband Entertainment. Westminster, CO: All Digital Publishing, 2004. A. Dhir. The Digital Consumer Technology Handbook. New York: Elsevier Press, 2004. B. Haskell. Portable Electronics Product Design and Development. New York: McGraw-Hill, 2004. A. S. Hoagland and J. E. Monson. Digital Magnetic Recording. New York: Wiley Interscience, 1991.

Chapter 2: Fundamentals of Hard Disk Drives S. X. Wang and A. M. Taratorin. Magnetic Information Storage Technology. San Diego, CA: Academic Press, 1999. E. D. Daniel, C. D. Mee, and M. H. Clark. Magnetic Recording. Piscataway, NJ: IEEE Press, 1999. F. Schmidt. The SCSI Bus and IDE Interface. Reading, MA: Addison-Wesley, 1997.

www.newnespress.com

286

Bibliography

Chapter 3: Fundamentals of Optical Storage J. Taylor. Everything You Ever Wanted to Know About DVD. New York: McGraw-Hill, 2004. T. W. McDaniel and R. Victora. Handbook of Magneto-Optical Data Recording: Materials, Subsystems, Techniques. Palo Alto, CA: Noyce Publications, 1997.

Chapter 4: Fundamentals of Flash Memory and Other Solid State Memory Technologies B. Dipert and M. Levy. Designing with Flash Memory. San Diego, CA: Annabooks, 1994. N. Gershenfeld. The Physics of Information Theory. New York: Cambridge University Press, 2000.

Chapter 5: Storage in Home Consumer Electronics Devices C. S. Swartz, Ed. Understanding Digital Cinema. New York: Elsevier, 2005. K. Jack. Video Demystified. New York: Elsevier, 2005.

Chapter 6: Storage in Mobile Consumer Electronics Devices L. Simone. If I Only Changed the Software Why Is the Phone on Fire? New York: Elsevier, 2007. J. Handy. The Cache Memory. San Diego, CA: Academic Press, 1993.

Chapter 7: Integration of Storage in Consumer Devices V. C. Hamacher, Z. G. Vranesic, and S. G. Zaky. Computer Organization. New York: McGraw Hill, 1984.

www.newnespress.com

Bibliography

287

N. Gershenfeld. When Things Start to Think. New York: Henry Holt and Company, 1999.

Chapter 8: Development of Home Network Storage and Home Storage Virtualization A. Kovalick. Video Systems in an IT Environment. New York: Focal Press, 2006. T. Clark. Storage Virtualization. Reading, MA: Addison Wesley, 2005. M. Farley. Building Storage Networks. New York: McGraw-Hill/Osborne, 2000. R. Perlman. Interconnections, 2nd ed. Reading, MA: Addison-Wesley, 1999.

Chapter 9: The Future of Home Digital Storage J. R. Pierce. An Introduction to Information Theory. New York, Dover Publications, 1980. D. Norman. The Design of Everyday Things. New York: Doubleday Currency, 1990. C. Anderson. The Long Tail. New York: Hyperion, 2006. J. Duato, S. Yalamanchili, and L. Ni. Interconnection Networks, An Engineering Approach. Piscataway, NJ: IEEE Computer Society, 1997. F. Bunn, N. Simpson, R. Peglar, and G. Nagle. Storage Virtualization, SNIA Technical Tutorial. Colorado Springs, CO: Storage Networking Industry Association, 2004. R. Cummings. Storage Network Management, SNIA Technical Tutorial. Colorado Springs, CO: Storage Networking Industry Association, 2004.

www.newnespress.com

This page intentionally left blank

Index

2K 63, 68, 93 3D 68 4K 63, 68 Abstraction 42, 217–218 Access time 65, 261 Active devices 11 Active tilt control 66 Adjacent cell disturb 87 ADSL—see asymmetric digital subscriber line ADSL cable 111 Aftermarket 177–178, 265 Ambient power 169 Ambient temperature 35–36, 277 American National Standards Institute 41 Ampex 108 AMPs—also known as Apple 8, 67, 121, 154–155, 177– 179, 281 Application integration ix, 20, 191, 198–200, 273 Application specific integrated circuit 57 Application specific standard products 89 Archival life 62 Areal density 25–26, 44, 46–49, 51, 59 ARM 168 Artifact 155

ASIC—see application specific integrated circuit ASSPS—see application specific standard products Asymetric digital subscriber line—see ADLS ATA 34, 40–41, 43–45, 50–51, 96, 114, 146, 176, 182–184, 192, 257–259, 263 Automatic metadata generation 127, 202, 248 Automobile ix, 3, 11, 38, 107, 131–134, 137–141, 170–172, 177, 213, 220, 244, 246, 256 Automobile entertainment 11, 107, 134, 137, 172, 220 Automobile navigation 133–134, 140, 171–172, 177, 246 Back up 10, 125, 129–130, 202– 203, 211, 229, 242, 244–245 Backseat entertainment 134 Bad block management viii, 42, 73, 85–86, 90, 104 Bandwidth 35, 66–67, 69, 94, 98, 124–128, 136, 151, 156, 163, 204, 207, 221, 239, 267, 269, 271, 276 Battery 10, 14, 37, 49, 132, 136– 137, 141–142, 144, 146–151, 153, 159–160, 165–167, 169, 177–179, 271 BD—see Blu-ray disk

BiCMOS—see bipolar complementary metal oxide semiconductor Bill of materials xiii, 7, 170, 175–176, 274 Bipolar complementary metal oxide semiconductor 74 Bit errors viii, 73, 83, 89, 95, 104 Blu-ray disc 59, 61, 65–67 Blue laser 8, 53–54, 57–62, 65– 71, 133, 162, 230, 232, 240, 252 Blue Onyx 37, 147 Bluetooth 37, 161, 169, 181, 266–268, 271 BOM—see bill of materials Buffalo 186, 211, 261, 283 Buffer 20–21, 35, 63, 139, 144, 149–150, 153, 192–196 Buyer’s remorse 62 Cable 3, 6, 19, 44, 108–109, 111–112, 114–115, 117–118, 122, 125, 128, 130, 141, 157, 177, 201–202, 204–209, 211– 213, 223–224, 229, 248, 258– 259, 266–272, 279–280 Cache memory 12, 15, 20, 286 Cache storage 205, 227–228, 233 CAGR—Cumulative Annual Growth Rate 46–47

www.newnespress.com

290

Index

Camcorder—see Digital video camera Camera 2–3, 10–11, 16, 37, 92, 95–96, 107, 114, 131–132, 135, 144–145, 158–167, 171– 172, 179–181, 189, 199, 204, 227–231, 237–238, 252, 262, 267 CBS 108 CCD 144–145, 159 CD 53–54, 57–58, 60–70, 98, 120, 133, 136, 156, 158, 170, 250, 269 CD-DA 54 CDMA—also known as code division multiple access CD-R 54, 203 CD-ROM 42, 54, 98 CD-RW 54 CE-ATA—see consumer electronics ATA Central processing unit 15 Channel hot electron injection 75–76 CHE—see channel hot electron injection Cisco 109, 207 CMC Magnetics 62 CMOS—see complementary metal oxide semiconductor Coaxial cable 122, 130, 204, 207, 211, 223, 269, 272, 279, 280 Codec—see encoder/deecoder Comcast 207, 279 Commercial content 2, 5, 123– 124, 127, 129–130, 141, 201, 223, 228, 230, 232, 239–240, 247–248, 253–254, 256, 264, 276 CompactFlash 98, 144, 262 Complementary metal oxide semiconductor 74 Compression 3, 66–67, 108, 126, 142, 155-156, 205, 221, 243 Consumer electronics ATA 44, 50, 115

www.newnespress.com

Content scrambler system 2–3, 10, 54, 60 Content sharing 3, 11, 124, 129, 180, 221, 225–226, 229, 232– 236, 244, 248–249 Content synchronization 220 Continuous wave 66 Controlled access 232 Controller 21, 23, 31, 35, 61, 83–85, 90–92, 94–98, 104– 105, 121–122, 135–136, 144, 159–161, 165, 183, 186, 191, 198, 216–218 Convergence device 2, 9, 167, 169 CPU 12, 15, 83, 109, 111–112, 135–137, 144–145, 185–186, 190, 192–197 Crucial technology 186, 261 Cultural heritage 254–255 Cumulative annual growth rate 46–47 CW—see continuous wave DAVE—see digital audio video experience DAC—see digital analog converter DAS—see direct attached storage Data beam 63 Data rate 15–16, 33, 35, 41–45, 51, 60, 64–66, 68–69, 113– 115, 126, 130, 148–150, 196, 204–207, 209, 215–216, 226– 227, 230, 268, 270, 279 Data retention 13, 93, 227, 243, 256 Data sharing 209, 225–226, 228, 233 Database 171, 184, 225, 240– 241, 247–248, 256 DataPlay 60, 70 David Rafner 108 DC/DC Buck converter 112 De facto standard 66 De-duplication 125, 244, 246, 276

Desktop computer 26, 38, 188, 212–213, 228–229, 231, 234 Digital analog converter 110 Digital audio video experience 37 Digital camera 131, 158–159, 163–164, 173, 229, 237–238, 267 Digital cinema 68–69, 286 Digital commons 223 Digital living network alliance 222, 264 Digital recording 108, 180 Digital rights management 97– 98, 222–223 Digital signal processor 137 Digital still camera 107, 158– 160, 189, 229, 231, 252, 262 Digital video camera 16, 160– 162, 229, 231, 252 Digital video recorder 3, 11, 16, 35, 38, 44, 51, 107–110, 113, 119, 129, 161, 177, 179–180, 199, 209, 251 Direct attached storage 5, 11, 40, 113, 228–229 Disaster recovery 5, 38, 128, 130, 209, 244, 248 Disc cartridge 60 Display technology 146–147 DLNA—see digital living network alliance Don Norman 171 Download 3, 11, 56, 61, 67, 95, 119, 124, 129, 133, 144, 155, 169, 205, 225, 226–227, 232, 234, 239, 252 DRAM—see dynamic random access memory Drive locking 115 DRM—see digital rights management Drone players 11 DVD 4–5, 8, 18–19, 53–55, 57– 62, 64–70, 109, 118–120, 133, 135–137, 140, 144, 156, 158, 162, 170, 179–180, 203, 205–

Index 206, 230, 232, 238–240, 264, 266, 268, 286 DVD+/−R 54–55 DVD+/−RW 55 DVD-Audio 5, 156, 205 DVD-RAM 55 DVD-ROM 54 DVD-Video 4, 54, 61, 205 DRV—see digital video recorder Dynamic Random Access Memory 12, 74 Echostar 207, 279 eCoupled 151 EEPROM—see electrically erasable programmable read only memory Electrical interface 25, 40, 51, 96, 98 Electrically erasable programmable read only memory 74 Electronic program guides 108, 119 Electrophoretic display 168 Elisa media center 128 Embedded 42, 73, 83–85, 104, 109, 113, 119, 122, 132–133, 147, 157, 168, 171, 193–194, 196, 259, 261, 262, 279 EMI 222 Encoder/decoder 110, 112, 143 Encryption 10, 128–129, 182, 186, 223, 255, 263, 266, 270, 275 Endurance 21, 77, 90–95 Entropic communications 207, 209 EPROM—see erasable programmable read only memory Erasable programmable read only memory 74 Erase/write cycles 77, 84–85, 92, 94 Erasure 76, 103 eSATA—see external SATA

Ethernet 40–41, 109, 112, 117, 124, 126, 130, 202, 204–207, 211, 214, 223, 227, 230, 232, 234, 268–270, 272 EVD 55 Evolution 26, 97, 100, 108, 127, 175, 197, 244–245, 261 External SATA 19, 114, 258 Fedex 239 Ferroelectric random access memory 102–103 Fibre channel 40–42, 219 Field effect transistor 74 FIFO buffer 63 File system viii, 35, 40–41, 50, 73, 86, 104, 184–185, 195, 199, 217–219 Firewall 124, 223 Firewire 11, 114, 161, 204, 211, 214 Firmware 9–11, 14, 36, 39, 85– 86, 176, 180, 182–183, 187, 190, 193, 196–197, 263, 274 First instance content 232 Floating gate 74–77, 86–88, 103, 281 Floating point gate array 112, 134 FM modulation 108 FN—see Fowler-Nordheim tunneling Format viii, 3–4, 8, 18, 33, 54, 57–58, 64, 66–68, 70–71, 74, 91, 95–99, 101, 105, 110, 137, 145–147, 155–156, 158, 162, 188, 205–206, 230, 240, 242– 243, 255, 260–261, 265, 274 Formatted 33 Fowler-Nordheim tunneling 75–76 FPGA—see floating point gate array FRAM—see ferroelectric random access memory Freevo 121 Front row 121 Fuel cell 150

291

Fujitsu 263, 282 Fusion 109, 203 GaN blue laser diode 65 GeeXboX 121 Global positioning system 137 GPS—see global positioning system Group participation factor 234 GSM—also known as global system for mobile communications HAMR—see heat assisted magnetic recording HANA—see high definition audio video network alliance Hard disk drive vii, 7, 13–17, 19–21, 23, 25–29, 31–42, 44– 46, 48–51, 59, 61, 70, 78, 83, 86, 90–91, 96, 103–105, 108– 110, 113–115, 118–120, 132– 134, 138–141, 145–146, 149–150, 152–158, 161–163, 169–172, 176–178, 180, 182, 184, 186–195, 197–199, 203, 205, 209–210, 214, 217, 219, 224, 229, 242, 250–254, 259, 261, 263–264, 274, 277, 282– 283, 285 HBA—see host bus adapter HD DVD 8, 55, 57–59, 61, 65– 67, 70, 156 HDD—see hard disk drive HD—see high definition HDTV 66, 68–69, 113, 119, 123, 126, 156, 165, 188, 205–206, 277 Head 21, 27–29, 32–35, 37, 39– 42, 46–47, 49, 51, 56–57, 62, 66, 107, 117–118, 153, 157, 161 Head-end storage 107 Heat assisted magnetic recording 46 Hermeticity 92 High definition 5, 35, 65, 66, 119, 162, 188, 205, 262, 268, 280

www.newnespress.com

292

Index

High definition audio video network alliance 280 Highway map 134 Hitachi global storage technology 282 HMA—see home media architecture Holographic read only memory 64 Holographic recording viii, 53, 62, 64, 69–70 Home automation 246–247 Home disaster recovery 38, 244 Home media adapter 107 Home media architecture 208 Home media center viii, 38, 107, 109, 117, 120, 123–129, 180, 198, 221–222, 245, 256 Home phone networking alliance 205 Home reference data 209–210, 224, 246 Home storage utility x, 225, 230, 240–246, 254–256, 273, 275 Home storage virtualization stack 218–219 HomePlug 204–206, 211, 279 Honeywell 108 Host bus adapter 219 HP 67, 283 HPNA—see home phone networking alliance HROM—see holographic read only memory HS-100 108 Humax 109 Hybrid storage vii, 1, 12, 17, 20, 22–23, 141, 187, 199 IBM 25–26, 263, 266, 269 IC—see integrated circuit IDE—see ATA IEEE 1394 11, 114, 122, 161, 204, 211, 214, 267, 280, 283 Imation 60 In-car navigation 134 Index x, 163, 184 Infineon Technologies 186, 261

www.newnespress.com

Information versatile disk for removable usage 138 Infrared laser 146 InPhase Technology 63–65 Integration ix, xiii, 5, 8–11, 20, 22, 31, 45, 51, 118, 128, 139, 147, 175–176, 183, 185, 187, 190–193, 195–200, 242, 258, 265, 273–274, 281, 286 Integrated Circuit—see IC International Microsystems Inc. 186, 261 Internet protocol 117, 272 iPhone 179, 281 iPod 8, 143, 154–156, 172, 178, 281 IP—see internet protocol IPTV 117, 119–120, 124–125, 205 iSuppli xiii, 68, 177, 179, 249–251 iVDR—see information versatile disk for removable usage JVC 158, 162 Keypad 111–112, 136, 142, 144, 146, 165–166, 194–195 Kingston Technology 186, 261 Landsat image 134 Laptop computer 20–21, 26, 36, 38, 51, 182, 228–229, 231, 283 Laser diodes 65–66 LCD 112, 136, 142, 144, 147, 159–161, 165–166, 178, 194–195 Length 9, 16, 42, 96–97, 114, 153, 190, 258 Lexar 186, 261 LG Electronics 60 License fee 118 Life-log 11, 131, 171, 173, 181, 210, 224–225, 227, 232, 240– 241, 248–249, 253–254, 256, 275 Linear density 38, 44, 49, 138 Linux 121–122

LinuxMCE 121 Location-based service 164, 181 LOCOS 100–101 Logical unit number 219 Longitudinal recording 29 Loss-less 3, 155–158 Lossy 3, 155 LSI 147 LUN—see logical unit number Macronix 88 Magnetic random access memory 12–13, 74, 78, 136 Magneto-optical 286 Management viii, 10, 42, 50, 73, 83–86, 90, 97–98, 104, 112, 117, 136, 142–144, 146, 150, 153, 159–160, 165, 169, 177– 178, 182, 193–195, 202, 211, 218–219, 222–223, 232, 242– 244, 246, 256, 266–267, 269, 275–276, 287 Mashup 247 Matrix Semiconductor 88 Maximum storage temperature 92 Maxtor 109, 115–116 MD—see mini disk Mean time between failures 35 Media over coax alliance 204, 279 Media portal 121 Memory stick 97–98, 146, 262, 266 MEMS—see microelectromechanical systems Metadata 10, 126–128, 143, 183–184, 202, 241, 244, 248, 256, 274–276 Metal oxide semiconductor 74 Metcalfe’s Law 233 Micro-electromechanical systems 46 Microprocessor 2, 12–13, 15, 23, 122, 184–185, 256, 263 MiniDisc 55, 60 MIPs 167 MirrorBit 88

Index MIT 151 MMC—see multimedia card Mobile consumer device 2, 11, 23, 131, 154, 168, 254 MOCA—see media over coax alliance Model home 228–229, 231 Moore’s law 58 MO—see magneto-optical MOS—see metal oxide semiconductor Motorola 109, 168, 207–208, 279 Moxi 109 MP3 player 3, 5, 84, 89, 95, 142, 149, 155–156, 164, 167, 172, 177–178, 188, 249–250 MPEG—see moving picture experts group MPEG-12 222 MPEG2 67, 113, 121, 206 MRAM—see magnetic random access memory M-systems 87, 186 MTBF—see mean time between failures Moving picture experts group— see MPEG Multicast 226 Multi-level cell 281 Multimedia card 97, 262 Multimedia object 68–69 Multiple level cells viii, 88–89, 104, 281 Multiple purpose device ix, 237–238 My life bits 171 My media system 121 Myth TV 121 NAND viii, 14, 16–17, 20, 22, 73–74, 76, 78, 80–87, 89–95, 97, 99–101, 103–104, 152– 154, 158, 165–166, 179, 186, 260, 281–282 Nano 143, 154–155 NAS—see network attached storage

National television standards committee—see NTSC Network attached storage 41, 113, 117, 129, 172, 218, 228–230 Network coding 221 Nichia 66 No permission limit 7 Node groups 234–235 Non-DRM 222 Nonvolatile memory 2, 4, 137, 175 NOR 12–14, 16, 74, 76, 78–80, 82–84, 93, 100, 104, 166, 281–282 Novalux 146, 148, 169 NPZ—see no permission limit NTSC—see national television standards committee Numerical aperture 58 Object-based storage 183–185, 199 Objects 68–69, 184 OEM—see original equipment manufacturer OMA—see open mobile alliance OMA—see optical mechanical assembly Open mobile alliance 222 Operating system 20, 32–33, 49, 111, 137, 145, 184, 196, 271–272 Optical absorption 59, 66 Optical mechanical assembly 61 Optical pickup unit 61 OPU—see optical pickup unit Original design manufacturers 50 Original equipment manufacturer 40 Pace 109, 115 PAL—see phase alternating line Panasonic 66–67, 207, 268, 279, 283 PAN—see personal area network PANS—see personal area network storage

293

Parallel ATA 43, 114 Parity bits 83, 104 PATA—see parallel ATA Patent 11, 25, 118 Patterned media 47–48 PC card 262 PCA 29, 31 PCI—see peripheral component interconnect PCMCIA card 96 PDA—see personal data assistant PDD—see professional disc for DATA-1 Peripheral 112, 177–178, 186, 245, 261 Peripheral component interconnect—see PCI 112 Peripheral enhancement 186, 261 Peripheral interface 163 Perpendicular recording 29, 46, 48 Personal area network 163, 270 Personal area network storage 37 Personal capacity 230–231 Personal content 5, 10–11, 38, 108, 123–126, 128–129, 201, 203, 210, 222–223, 225, 227, 229–230, 232, 239, 242, 247– 248, 253–256, 264, 274–275 Personal content creation device 11 Personal data assistant 248 Personal life recorder 171, 189 Personal map 240–241 Personal media player 3, 144– 146, 157, 178, 190, 199 Personal memory assistant 225, 240–241, 249, 254–256, 275 Personal petabyte 210 Personal video recorder 36, 108 Petabyte 126, 171, 210, 232, 248 Phase 13, 42, 54–56, 60, 191– 197, 199, 244–246, 256 Phase alternating line—see PAL Phase change random access memory Philips 58, 60, 62, 66–67, 264

www.newnespress.com

294

Index

Photographs 2, 4, 20, 37–38, 92, 95–96, 120, 128, 129, 141, 157–158, 180, 209, 220, 238 Physical content distribution 53, 68, 225, 239–240 Pixel 68, 158–160 Plain old telephone system 205 PLR—see personal life recorder PMA—see personal memory assistant PMP—see personal media player PNY Technologies, Inc 186, 261 Pocket projector 146, 169 Point-to-point 225 Polycarbonate 66 Portability 154–155, 259 Portelligent 13, 145, 160, 162, 166 POTS—see plain old telephone system Power 2, 12–16, 20–21, 31, 33– 39, 49–50, 61–62, 66, 70, 83, 103, 105, 112, 114, 122, 130– 131, 137–139, 141–143, 146– 153, 156, 159–160, 166, 168–169, 172, 176–178, 191– 192, 195–196, 203–206, 216– 217, 223, 237, 258, 261, 267, 270–272, 274, 277, 279, 282 Power on hours 34, 62, 70 PRAM—see phase change random access memory Pre-amplifier 29 Printed circuit assembly 29 Privacy 10, 123, 125, 128, 167, 181, 186, 201, 222, 239, 241, 246, 255–256, 264, 275 Professional disc for DATA-1 55 Prosthetic memory 240 PVR—see personal video recorder QoS—see quality of service Quality of service 117, 267, 270 Quickview 116 Radio Frequency 270 Radio receiver 143, 164

www.newnespress.com

Radio Shack 207, 279 RAID—see redundant array of independent disks RAMAC 25–26 Read duty cycle 35 Read head 29 Read only 54, 58–60, 66 Read only memory 13, 64, 74 Read/write channel 29, 61, 65 ReadyCache 49 ReadyDrive 20, 49 Red laser 18, 58–60, 67, 70 Redundant array of independent disks 219 Reed’s law 232–233 Reference laser beam 63 ReplayTV 108 Repository 125, 127–128, 209, 220 Return rate 62 Revolution 108, 127, 261 Rewritable 26, 54, 59, 66, 140, 171 RF—see radio frequency Ritek 60, 62 Roadmap 99 Robson—see turbo cache ROM—see read only memory Rotary actuator 39 RPM 41, 49 Rules vii, 1, 9–10, 22, 37, 181, 191, 238, 274 SACD 54 Saifun NROM 88 Samsung 66–67, 103, 155, 166, 186, 260–261, 264, 268, 277– 278, 281, 283 Samsung Semiconductor Inc. 186, 261 SanDisk 87, 92, 97, 119, 186, 260–261, 282 SAN—see storage area network Sanyo 60, 62, 67, 166 SASI—see Shugart Associates system interface SAS—see serial attached SCSI SA-STI 100–101

SATA—see serial ATA Satellite 3, 108–109, 115, 117, 119, 125, 128, 137, 140, 157, 206, 209, 212–213, 229, 246 Satellite radio 119 SaveTV 121 Scientific Atlanta 109, 115, 117, 207–209 SCSI 40–43, 114, 183–184, 188, 219, 263, 285 SD—see standard definition SDTV 54, 66, 113, 123, 126, 156, 227 Seagate 29, 37, 47, 109, 115– 116, 147, 211, 258, 283 SECAM—see Sequential color with memorySecure data card Secure video processing 222 Self monitoring and reporting technology 34, 191 Sensor 27, 29, 30, 37, 48–49, 108, 134, 138, 144, 159–160, 162, 165, 246, 271 Sequential color with memory— see SECAM Serial ATA 43, 114, 146, 258–259 Serial attached SCSI 43 Service model 206–207, 209 Service providers 107, 113, 128, 130, 181, 201, 207, 224, 249 Servo control 65 Set-top boxes 11, 109, 116, 125, 171, 180, 207–208, 264, 281 Sharp 62, 67, 279, 282–283 Shock 14, 16, 21, 35–37, 39, 45, 49, 92, 105, 132–133, 138, 140, 152–153, 278 Shugart Associates system interface 42 SiBeam 268 Single level cell viii, 22, 73, 86–87 Single purpose device 225–228, 232–233 Single spouse permission limit 7 Single use device 226–228

Index SLM—see spatial light modulator SMART—see self monitoring and reporting technology SmartMedia 85, 96–98, 262 SMT—see surface mount technology SNIA—see storage networking industry association Social network ix, 3, 171, 180, 221, 223, 225, 239, 253 SoC—see system on chip Solid state drive 134, 138, 176, 277 Sony 55, 60–62, 66–67, 97–98, 145, 260–261, 265–266, 268, 283 Spansion 88, 261, 282 Spatial light modulator 63 Specific gravity of 1 36 Spherical aberration 60, 66 Spindle motor 27, 39 SRAM—see static random access memory SSPL—see single spouse permission limit Stacking die 89 Standard applications 2, 119 Standard consumer application 9, 274 Standard consumer function 179–181 Standard definition 18, 35, 54, 66, 109, 113, 118, 126, 221, 227 Start stop cycle 35 Static consumer device 11, 18, 181 Static random access memory 12 STB—see set-top box Storage area network 41 Storage hierarchy 1, 13–19, 131, 176, 187, 242 Storage networking industry association 243, 275 Storage virtualization 201, 217– 220, 256 Streaming 35, 63, 82, 108, 113, 119, 126, 136, 150, 156, 176,

181–182, 205, 207, 227, 239, 268 Surface emitting laser 146 Surface mount technology 57 S-Video 110, 145 SVP—see secure video processing Synchronization 140, 169, 181, 211–212, 220, 244, 248 SYSCON—see system controller System controller 61 System on chip 31, 193–195 T10 183 T13 176, 182–183, 257–258 Tape 13–14, 16–18, 25–26, 42, 46, 48, 62, 65, 158, 162–163, 217, 229 TAR—see HAMR TDMA—also known as time division multiple access Teardown xiii, 142–145, 159– 162, 166 Television 3, 53–54, 66, 108, 117, 119, 180, 208, 237–238, 264, 267, 271 Temperature 16, 33–39, 70, 92, 132–133, 138, 140, 153, 217, 277 Terabyte 26, 118, 140, 171, 232, 248, 254 Texas Instruments 110, 112, 136, 142, 144, 159, 165, 261 Thermal stability 47 Thermally assisted recording— see HAMR Thomson 109 TiVo 108–111, 117–118 TMR—see tunneling magnetoresistive reading Topfield 109 Topographic base map 134 Toshiba 29, 47, 66–67, 96, 100, 207, 258, 266, 268, 277–279, 282–283 Total product cost 10, 22, 176– 178, 198 Track density 38, 138

295

Trek 186 Trick modes 113, 126 Triple play 110 Trusted computing group xiii, 128, 182–183, 185, 262 Tunneling magnetoresistive reading 47 Turbo Cache 49 TVedia 121 TV—see television U3 98, 99, 186, 198, 261 Ucentric Systems 208 UDO 55 UFDA—see USB flash drive alliance Ultra high definition TV 119 Ultra-HD 156, 221 USB 11, 15, 20, 49, 84, 99, 101, 109, 114–115, 117, 119, 122, 137, 141–142, 144, 146, 160– 161, 165, 177, 185–186, 188, 198–199, 211, 214, 239, 258– 261, 283 USB flash drive alliance 185, 198, 260 USB TV 119 User interface 111–112, 137, 143, 145–146, 159, 166, 218, 238, 241, 273, 280 Utilization 230–231, 253, 270, 275–276 VCD 54 VCR—see video cassette recorder Vehicle information and communications system 134 VHS 18, 108 Vibration 14, 17, 21, 35–38, 92, 105, 132–133, 138, 140, 152, 153, 278 VICS—see vehicle information and communications system Video 2–5, 10–11, 15–16, 18–19, 21–22, 35–38, 44, 51, 53–54, 60–61, 63, 65–68, 70–71, 82, 107–114, 117–124, 126,

www.newnespress.com

296

Index

128–129, 131–136, 138, 140– 141, 143–147, 150, 156–158, 160–164, 166–167, 170–172, 177–180, 182, 184, 199, 203– 206, 208–210, 220, 222–223, 227–239, 244–245, 250–252, 267–268, 270, 280, 286–288 Video cassette recorder 108 Video on demand 113, 121 Virtualization 33, 59, 201, 203, 217–220, 230, 232, 243–244, 246, 256, 275–276, 287 Vista 20, 33, 49, 50, 121 Vladimer Paulsen 25 VOD—see video on demand

www.newnespress.com

Volatile memory 2, 21, 74, 137, 175 Volume manager 218 Wear leveling 73, 84–86, 91, 93– 95, 104 Weight 38, 141, 146–147, 152– 154, 185, 241 Western Digital 43, 115, 283 Width 9, 16, 33, 96–97, 153, 175, 182, 187 WiFi 205 WiHD—see wireless HD Windows media center 121 Wireless HD 268

Wireless power 150, 172 Witricity 151 Write duty cycle 35 Write head 29 Write once 54, 64, 66 X-box 360 120 Xbox media center 121 xD-Picture Card 85, 96–98, 262 YouTube 239 YPrPb component video 110 Zoning 33, 219