GLG News by Thomas Coughlin

Implications: § By some estimates there are over 50 companies currently offering products that they call solid state drives (SDDs)  §  Many of these products have been designed hastily and some of them show no improvement in benchmarking tests vs. hard disk drive metrics  §  There will be a significant consolidation and culling of the SSD players over the next 1-2 years  §  It is likely that SSD consolidation will mirror that of HDDs in the 1980’s

Analysis: At the Flash Memory Summit (August 12-14, 2008) in Santa Clara, CA it became clear that there are a great many companies from all over the world that are offering flash based SSDs.  By some estimates there are over 50 of these companies.  Some of these companies are relabeling other manufacturer’s products but many are buying flash memory and controllers, writing some firmware and introducing their own products.  Some of these companies are specializing in the enterprise (high performance) SSD space and some are specializing in embedded and industrial markets.  Still other are hoping that flash memory will play a major role in laptop and mobile computer applications.  

The number of SSD companies reflects a perceived “gold rush” of people hoping to stake a claim on what many feel could be a flash memory bonanza.  There have been several enterprise flash memory SSD products introduced and integrated in to storage systems by companies such as EMC.  Computer manufacturers such as Dell and Lenova have made significant commitments to flash-versions of laptop models and in some cases to SSD only laptops.  Total volumes of SSDs for all of these applications have been modest but many projections predict robust growth of these storage devices.  

Many of the companies making SSDs will probably not be around in a few years time, a victim to consolidation and bankruptcy.  It is likely that the SSD industry will follow a trend that was seen in the HDD industry.  In the early 1980’s there were over 80 companies manufacturing hard disk drives.  By the 1990’s there were fewer than a dozen and today there are only 6-7 companies making hard disk drives.  When a market first emerges many people feel that they can enter the market and compete even if they don’t have a clear differentiator from their competition or they have supply chain problems.  As the unit volume in the industry increases prices drop and many of the competing companies are no longer competitive and either must merge with other companies or go out of business.  

The number of SSD companies will probably decline by 30% over the next three years.  Companies that don’t manufacture their own flash memory will be under a cost disadvantage as the prices of SSDs go down and face increasing price competition.  Likewise SSD companies that don’t have their own SSD controller technology will be at a significant cost as well as performance advantage.  Companies that have good controller IP as well as access to flash memory manufacturing will be most able to weather the early days of SSD competition.  This will be especially true in the rough and tumble world of laptop and mobile internet devices.  

Some of the companies that will probably survive the initial market culling include Samsung, Intel, SanDisk, Toshiba and Micron.  Companies in the enterprise market that have sophisticated controller IP may survive longer even if they don’t have their own flash factories since these products have higher margins.  The current surplus of flash memory will help these companies keep their costs down as well.   

As flash increases in volume and as a clear space in the storage hierarchy is created for SSDs (and following a similar trend to the early HDD industry) there will probably be less than 15 players in 10 years time.  Most unit volume by then will probably be held by 5-6 players with the balance being niche market manufacturers.

Implications: * There are rumors that Western Digital is working on a 20,000 RPM Raptor drive  * Other HDD companies such as Seagate, Fujitsu and others are also working on improving rotational as well as track change drive latencies  * SSDs are starting to compete against HDDs for performance oriented applications, one of the more lucrative, higher margin parts of the HDD market  * HDD companies will need to respond to the incursions of SSDs into the enterprise and high performance space by using new technologies

Analysis: Solid State Drives (SSDs) are being incorporated into storage systems to provide faster data in enterprise and performance applications.  EMC and other storage companies have announced storage systems using SSDs for high performance applications.  These flash based SSD products from companies such as STEC are targeted to compete against the much more expensive DRAM based SSDs used for very transaction oriented applications.  By reducing the price of fast transaction products this capability can be used more widely in the market.  

In addition to competing and extending the market previously held by DRAM based SSDs these enterprise flash based SSDs will also compete against hard disk drives used in applications where the storage capacity of the drives is sacrificed in order to achieve high aggregate data throughput in a storage array.  Such systems use many hard disk drives in an array to achieve fast processing speeds.  A single flash based SSD in these applications may be able to replace several HDDs in terms of data throughput and thus provides a lower cost alternative.  If the primary application is read intensive with lesser number of writes such applications will encroach on one of the traditional markets for high performance enterprise hard disk drives.    

It is estimated that this could affect the market demand for as much as 10% of the high performance enterprise hard disk drives (Fibre Channel and SAS interface disk drives) within the next few years.  HDD companies that could be impacted by this shifting demand for some of the highest performing (and highest margin) drives are Seagate, Fujitsu and Hitachi.  

Another application for high performance hard disk drives is gaming applications and higher performance SATA workstations and arrays.  This is an area that Western Digital has developed with its 10,000 Raptor SATA hard disk drives.  Hard disk drives offer less expensive bulk storage than SSD and will for many years to come, but as in the enterprise high performance market there are applications that need faster data transactions.  For these applications, if the storage capacity required is not so great, a SSD may be a good option.  

The disk drive companies, including Western Digital, are looking at ways that they can boost the speed of their highest performing HDDs.  Even for the high speed enterprise HDDs the maximum rotation rate of the disks is limited to 15,000 RPM.  A move to 20,000 RPM would decrease the rotational latency (the speed it takes to access data on a track) by 33% and thus increase the data rate available to write and read information on the disk drive.  Likewise other changes to the disk drive can reduce the track to track latency.  By moving to higher performing technologies HDD companies can combat some of the incursions of SSDs into their markets.  

HDD companies have not done much to improve mechanical drive performance for many years, leaving them vulnerable to competition to solid state storage products that don’t have moving parts.  They now need to make significant investments in higher performing products in order to maintain and expand their penetration in these growing markets.  The world is not getting any slower!

Implications: * Nintendo and InPhase are apparently doing some joint work on holographic storage  * InPhase has been pursuing a holographic storage system (primarily for write once archive applications) for several years  * Any holographic storage product likely will not be in a Nintendo game system for several years  * It is likely that Nintendo will need some other storage device such as a hard disk drive for its current Wii storage limitations

Analysis: InPhase Technology in Longmont Colorado has been working for many years on holographic storage on an optical disc, using technologies originally developed at Bell Laboratories.  Although the company has made several commitments for delivery of 300 GB write-once discs for archive applications including professional entertainment content archiving over the last three years there has been no official shipment of volume product.  At this point it is not clear when the company will begin to ship product or how long the start-up’s venture funding will last.  

Holographic storage has never been made into a commercial product.   There are many components in an holographic storage system that have never been produced in the volumes and in the space required for commercial or even consumer applications.  Shrinking these components and a getting reliable storage system has taken more effort than originally estimated.  As is often the case, pioneering companies may develop a technology but not be able to introduce it before their funding runs out.  It remains to be seen whether InPhase will be able to cross this particular chasm.  

It was recently revealed that Nintendo has a Joint Development Agreement with InPhase Technology on holographic storage technology.  Nintendo is a leader in electronic game technology, making the very popular Wii game system.  Currently announced InPhase technology would provide a 300 GB cartridge but this was for write once applications.  A physical distribution media with InPhase is also possible.  InPhase has said that their holographic technology could deliver storage discs with capacities as high as 1.6 TB.  Such high capacity would enable very high resolution game content as well as game storage.

Implications: * SanDisk recently revealed a proposal for creating a single metric representing the expected lifetime of a solid state drive  * The SanDisk metric is called “Longterm Data Endurance” or LDE  * The SanDisk proposal comes at a time that several organizations such as IDEMA and JEDEC are discussing standards for SSD reliability  * SNIA, the Storage Networking Industry Association, is also launching a solid state initiative that will explore the evolving use of solid state storage  * These are important developments for sustainable growth of SSDs in the enterprise and computer markets

Analysis: Sandisk released details on its proposed standard metric for solid state drive (SSD) expected lifetime.  This metric is called Longterm Data Endurance (LDE).  SSDs are subject to wear at the flash memory cells that limit the number of times that these cells can be written.  Flash memory devices such as SSDs deal with this wear using special algorithms that run in the flash memory controller, called wear leveling.  Wear leveling spreads the wear around the flash memory cells to prolong the life of the memory cells.  Wear leveling is even more important for multilevel cell flash where multiple bits are stored on each cell since this sort of flash memory cells wear much faster.  With the proper wear leveling and excess capacity for memory management and space sectors flash memory can be made to last for quite a while, depending upon the usage.

At about the same time that SanDisk introduced this proposed standard the IDEMA reliability group that created standards for HDD reliability has begun to work on creating new standards for SSD reliability.  Similarly JEDEC, another standards group that focuses on solid state memory technology has begun an effort to create metrics for endurance, reliability, performance and data lifetime for flash memory devices.  SNIA, the Storage Networking Industry Association is initiating a flash memory initiative of its own that looks at new uses for solid state storage in storage system architecture.  Clearly there is much work going on in this area.

It is good that standards bodies are starting to look at creating verifiable standards for SSDs.  There has been much hype over the last few years on SSD performance and reliability and many people have found that the actual increase in performance and reliability are no better than many hard disk drives despite the higher prices.  The results conflict based partially upon the test methodology and measurement methods used, so clearly there is need for agreed upon measurement methods and reliability metrics.  This is what standard organizations can create.  On the other hand it needs to be determined what each of these various groups will be doing regarding SSD standards.  Competing standards will not clarify the situation for designers or users.

Moving from relatively well understood chip level reliability metrics to solid state drive reliability metrics is complicated by the various features and algorithms running on various flash memory controllers that implement wear leveling and other flash memory controller functions.  The LDE metric proposed by SanDisk measures the total number of Terabytes a solid state drive can write in its life.  The standard is based in part on Bapco standards for computer performance metrics.

It remains to be seen whether the SanDisk endurance metric will be used throughout the industry but there is a clear need for the various potential standards groups to work together to create a meaningful SSD reliability metric.  This metric should be capable of comparing various SSDs as well as the endurance of SSDS compared to HDDs.

Implications: * JEDEC is a memory standards body that is very active and respected in the electronics industry  * JEDEC has formed a committee to develop standards for NAND flash solid state drives  * The JEDEC committee will be co-chaired by Micron and Seagate  * Seagate expects to announce their first SSD product for the enterprise market by the end of 2008  * The JEDEC standardization effort is part of a standards movement for SSDs that involves a newly formed SSD Alliance, IDEMA and the ANSI T13 committee

Analysis: Solid state disk drives are starting to appear in systems.  In the computer market these products are targeted at the enterprise (high performance) or laptop markets.  The SSD products for each of these products are very different because of their difference performance and reliability requirements and flash memory will play different roles in these two markets.  Solid state drives have gotten a lot of press and some wild claims have been made concerning power savings, reliability and other important characteristics with little actual supporting data.  

In the past few months there have arisen a number of standard efforts to deal with difference between HDDs and SSDs, ways to test for performance and reliability, reliability specifications and form factors.  Some of these standards activities are overlapping and jurisdiction for the various standards will have to be ironed out between the various groups.  

JEDEC Committee JC 64.8 will be responsible for defining new form factors, leveraging existing interface standards, specification of environmental requirements, electrical quality, reliability and durability methods and procedures.  The IDEMA reliability standards group, lead by Jon Elerath from NetApp has gathered a group of HDD and SDD vendors to define new reliability specifications.   

The SSD Alliance (formed by a number of Taiwan based companies) will release its initial specification for SSD testing in July 2008.  The SSD Alliance plans to pull in chip makers as well as PC vendors into the standards effort.  The ANSI T13 committee is responsible for the ATA interface specifications.  This committee has created some new standards for SSD memory management based upon different memory requirements for SSDs vs. HDDs.  

At the just completed Memcon conference SanDisk announced new standard proposals for SSD endurance testing--Longterm Data Endurance (LDE) and performance.   It will be interesting to see what becomes of these standards and where this standard-making activity will be done.  

Standard setting is an important requirement in bringing new products to market.  Without standardized measures, commands and testing it is hard to compare SSD products with each other and with HDDs.  This makes it difficult for vendors to incorporate SSDs into computer and other products and make it difficult for the SSD vendors to deliver consistent products and to anticipate their performance and reliability in the field.

Implications: * Seagate announced their Barracuda 1.5 TB disk drive, the first announced disk drive with more than 1TB capacity  * Seagate also announced their first 500 GB 2.5-inch drive products with 5200 and 7200 RPM  * The 3.5-inch 1.5 TB product will begin shipping in August 2008 while the 2.5-inch 500 GB products will ship in Q4 2008  * Separately Hitachi said that 5 TB HDDs will be possible by 2010

Analysis: HDD storage capacities continue to expand.  Last year saw the introduction of 1 TB 3.5-inch drives by Hitachi, followed by all the other 3.5-inch disk drive manufacturers.  This year Seagate has announced 1.5 TB capacities in this form factor.  The Seagate Barracuda drive supplies 1.5 TB on 4 disks (375 GB per 3.5-inch disk).  This is the highest capacity ever announced in a hard disk drive.  The 1.5 TB Barracuda runs at 7200 RPM. 

1.5 TB storage capacity supplies a lot of storage for computer, consumer and enterprise applications.  I expect to see additional offerings of 1.5 T and maybe even up to 2 TB storage capacity in the next few months.  A recent article also says that Hitachi Global Storage Technology expects 5 TB disk drives by 2010.  To put 1.5 TB in perspective, this is enough storage capacity for 319 DVD quality movies and about 60 Blu-ray (at 25 GB each) movies. 

Unlike in the past when HDD areal densities grew at greater than 100% annually, the more moderate growth of 40-50% on average per year is lower than the growth in storage demand.  As a consequence I don’t believe that there will be any significant negative impact on disk drive demand or component count (heads and disks).  Instead I see continual growth in both shipping storage capacity and hard disk drive storage capacity—especially with advances in media and head technology foreseen over the next 5 years.

Higher HDD areal density drives enable new applications and provide a significant economic advantage in $/GB compared to other storage technology, such as flash memory.  Most hard disk drives are used for mass storage applications and over the last 5 years the demand for digital storage has increased as the price of storage has declined. 

I have come up with a couple of metrics to represent this growth in demand over this period.  These are:  1) storage capacity shipped increased by 2.1 Exabytes per every percent drop in the $/GB  of drives and 2) HDD units shipped increased by 3.2 million units for every percent drop in the $/GB of drives.

Storage demand, and hence disk drive demand is very elastic.  As the price of storage capacity declines it enables increasing numbers of applications using ever higher amounts of storage capacity.  I estimate that by 2012 annual shipments of hard disk drives could exceed 1 billion units per year (up from just over 500 million units in 2007). 

Not only are applications directly using hard disk drives benefited by their growing capacity but so are many other applications using other types of digital storage.  For instance flash memory is used for many mobile playout devices as well as content capture consumer devices such as still and video cameras.  In all these cases the flash memory is temporary storage with long term storage on hard disk drives. 

As larger hard disk drives become available the support for these other devices increases and demand for both the hard disk drives and the flash memory in these consumer products increases.  Furthermore as all this content increases so does the value of backing up content, again on hard disk drives.  Large hard disk drives are good for storage growth, all sorts of storage growth.

Implications: * USB 2.0 provides external data rates of 480 Mbps sustained.  eSATA, an external version of the SATA interface currently provides data rates of 3 Gbps sustained  * Next generation USB will have 5 Gbps and next generation eSATA will have 6 Gbps data rates * Recently several laptop and external storage companies, many based out of Taiwan, have introduced so-called 2-in-1 connectors that support both USB 2.0 and eSATA connector devices Using the space saving 2-in-1 connections customers can use popular USB 2.0 devices or very fast eSATA connections  * This will assist in the growth of eSATA external storage devices

Analysis: USB devices have become ubiquitous.  Thumb or flash drives seem to proliferate like mice in a cheese plant.  Everyone is giving these things away, often with catalogs or other company information on them.  These devices have become the de-facto way to transfer files and data from one electronic device to another.  Likewise for backup and other higher capacity storage applications external hard disk drives with the USB interface are very common.  USB interfaces are found on almost all commercial and consumer electronic devices in the market.  The current USB 2.0 interface provides data rates up to 480 Mbps.  

eSATA is an external version of the SATA interface that is used in desktop and laptop computers as well as in many consumer devices and an increasing number of disk drive storage arrays.  The eSATA interface provides data rates up to 3,000 Mbps (3 Gbps), roughly 6X faster than the USB 2.0 interface (actual data rate differences are different due to the overhead used in each interface standard as well as the peculiarities of the particular application).  The high data rates of the eSATA interface provide an ideal interface for video editing, backup and other data intensive applications where the a slower interface speed would limit the performance of the user experience.  A faster data rate allows high resolution video frames to be streamed off the external storage device without interruption.   

The eSATA interface has such a high data rate that several streams of content can be handled at once across the interface (the number of streams depends upon the resolution and compression of the content).  For this reason most new Digital Video Recorder (DVR)-capable set top boxes used in cable and satellite TV as well as stand alone DVRs now include an eSATA port that can be used to expand the internal storage capacity of the DVR device.   

USB 3.0 has been announced for implementation in the next several years with a data rate of 5 Gbps but and would be a new interface with backward compatibility to the existing USB 2.0 connectors.  The next generation eSATA would provide 6 Gbps data rates and will probably be available before USB 3.0 connectors.  Such high speed interfaces can provide even greater capability to handle rich content and multiple stream content.  

The eSATA port has required special connectors to implement and this can add to the cost of implementation and has limited the number of devices that provide this very fast interface.  The recent rash of devices using the so-called 2-in-1 connectors that support both USB 2.0 and the current generation of eSATA.  Although not an official standard interface this hybrid connector allows a low cost, space saving capability that gives users the choice or ubiquitous USB or blazingly fast eSATA external storage devices.  At the recent Computex conference in Taiwan there were a plethora of laptops and other popular consumer devices sporting these hybrid connectors.  

The 2-in-1 USB/eSATA connector looks to be a hit in the computer industry and I expect these combination interfaces to proliferate.  The growth of these connectors supporting both USB and eSATA will also help the external eSATA interface get more traction in a broader market.  This will help to drive the adoption of eSATA and promote growth of companies that support eSATA including Silicon Image, Marvell, Oxford Semiconductor and Intel.

Implications: * Oerlikon sold its disk drive sputtering assets to Intevac and exits the hard disk drive business  * Oerlikon, in a former incarnation as Balzers used to be one of the major players in the disk sputtering equipment market but had lost most of its business over the last 10 years  * Despite making efforts to introduce new products in the last few years Oerlikon placed a distant third to Intevac and Anelva in the media sputtering equipment market  * Oerlikon had also been sued by Intevac for violation of patents on disk handling in sputtering systems  * The merger of Oerlikon’s disk equipment business into Intevac is only the latest development in capital equipment consolidation for the industry and certainly not the last

Analysis: Back in the mid-1990’s Balzers was one of the major manufacturers of disk sputtering equipment with equipment at many major disk manufacturers at the time including IBM and WD Media.  At the end of the 1990’s and early 2000’s Balzers underwent significant changes in focus (as well as name—they went through at least 4 name changes in about 5 years).  During this period the company lost its focus on the disk drive production equipment industry. 

The slow down during the dot-com bust combined with aggressive areal density increases reducing the average number of disks per disk drive left the industry with lots of surplus disk manufacturing equipment.  As a consequence there were several years with little or no new equipment purchases.   During the slow down companies such as Intevac and Anelva had to survive on equipment upgrades and maintenance while they tried to create some diversification in their customers. 

Anelva was acquired and became part of Canon and Balzer (and its many named descendents) ceased producing new products for the industry for several years.  When the market returned for more disk production equipment driven by disk drive unit volume growth, the cessation of component count declines in disk drives and the conversion to perpendicular recording Intevac and Anelva once again saw a renewed market for their equipment.  

In 2006 Oerlikon (the latest manifestation of former Balzers) introduced a new line of sputtering equipment but by then they had few if any customers remaining from their old Circulus line of sputtering machines.  The new equipment had a few sales to Samsung and some other companies for research purposes but it was never adopted for volume production by any of the disk manufacturers.  In addition, in 2007 Intevac filed a lawsuit for patent violation against Oerlikon.  

The disk drive capital equipment industry is extremely cyclical, especially for expensive, long lead time capital equipment like disk sputtering equipment.  While the equipment can take perhaps nine months from initial order until the equipment is installed at the customer and starting to build product, the market for disk drives can change in weeks and often undergoes long term spells of either intense sales or utter stagnation.   Thus 2006 was a great year for disk sputtering equipment sales as unit volume growth and the conversion to perpendicular recording drove sales of new sputtering equipment.  However by the second half of 2007 and in 2008 the market has declined considerably.  

Consolidation in the disk drive and component capital equipment industry is on-going and the Intevac purchase of Oerlikon’s disk sputtering technology should help the remaining vendors improve their overall profitability when the market picks up again—probably in 2009 with increased demand and the introduction of discrete track recording media.   

Stronger vendors will be better able to support the research and development as well as the cost of manufacturing these ever more complex machines.  As pointed out in Coughlin Associates’ "2008-2013 Capital Equipment and Technology Report for the Hard Disk Drive Industry" (www.tomcoughlin.com) total capital spending between 2007 and 2013 will be about $35 B to support cumulative revenue that could reach $313 B.  The price of disk sputtering equipment including patterned media manufacturing capability could rise to $10-15 M each.

Implications:  Seagate has stated that they will introduce a SSD device by the end of 2008 This is expected to be an enterprise level SSD to help protect the company’s position in the enterprise storage market There are several rumors that Seagate is looking at acquisitions to enable its SSD effort Intel is looking at extracting itself from its unprofitable joint flash venture with Micron Technology Seagate acquisition of a source of flash memory and flash memory IP could help it establish itself as a player in the emerging SSD market

Analysis: Seagate has stated that they will introduce their first solid state disk drive (SSD) product before the end of 2008.  This is likely to be an enterprise SSD that can help to maintain Seagate’s dominance in the enterprise storage market.  Although the initial market impact of enterprise SSDs may be to expand the market and take away market share from DRAM-based SSDs used for input/output (IOPS) intensive applications these products could eventually provide competition with some high performance hard disk drive applications which are IOPS-intensive.

The rapidly declining price of NAND flash memory and the increasing complexity of flash memory controller to enable better performance and reliability may help the technology make inroads into the high performance market.  It should be noted that the performance and reliability requirements for digital data storage applications are much more demanding than for consumer music, video and photograph applications where NAND flash is more often used for mass storage.  For computer applications there are significant complications in data integrity and error correction as well as interface and communication requirements that consumer flash applications have not had to deal with.

For this reason SSD companies are hiring hard disk drive electronics engineers to help them develop the controller firmware and hardware to give these products the capability of performing in the demanding data storage market.  This is also a reason why companies such as Seagate, already familiar with data storage drives, could do well in this market.  This is also a reason why Seagate may want to acquire IP and resources associated with flash memory companies.  There are rumors that Seagate is talking with STEC (who it has sued for using HDD interface technology in its SSD devices) and Intel/Micron Technology’s flash memory joint venture.  Combining internal device production, flash IP and Seagate’s hard disk drive channel and interface experience could help Seagate maintain their market position in the enterprise market and also give additional credibility to SSD in the enterprise market. 

Despite the promise of using SSDs in enterprise applications it is likely that unit sales of SSDs for many of these applications will not be high for some time.  One of the biggest applications would be to replace a large array of high performance hard disk drives with a few SSDs giving the same IOPS performance.  Thus the total unit sales of these SSDs will substract from HDDs but add only a fraction of the number of SSDs in their place. 

There is an additional additive market that would replace and expand expensive very high performance DRAM-based solid state disk drives that have been in the enterprise market for many years with NAND SSDs.  In any case SSDs in the enterprise will be expensive and unit demand will likely not be significant for some years to come even if the devices achieve technical success.  However these devices will be high profit margin devices and could add to participating company’s bottom line.

Implications: * Hitachi president Kazuo Furukawa said that Hitachi wants to surpass Seagate’s HDD business profitability  * Furukawa-san said that Hitachi has the latest technology required to be competive in the HDD business  * He also said that there is synergy between Hitachi’s computer business and its HDD business  * In the last two quarter Hitachi has reported profitability, although no where near the levels of Seagate and Western Digital

Analysis: Hitachi’s hard disk business, Hitachi Global Storage Technology (HGST) lost money from 2005 until the last calendar quarter of 2007.  It is only in the last two quarters that the company has returned to profitability.  This profitability was achieved by laying off workers, consolidating manufacturing sites and pairing down some smaller and less profitable HDD programs, such as their 1.8-inch disk drive programs. 

The president of Hitachi says that his goal is to surpass Seagate’s profitability but there are many challenges that HGST will have to surmount to achieve this goal.   Hitachi has had a history of management and cultural conflict, left over and never properly resolved after the merger of Hitachi and IBM’s HDD division to create HGST. 

This has resulted in many program misfires, in particular HGST lost most of its market share in the enterprise HDD space to Seagate and Fujitsu as a result of problems getting generations of enterprise HDDs to market.  Back in the 1990’s IBM was the market leader in high performance HDDs.  These management issues will need to be addressed and resolved if HGST hopes to have a chance to permanently improving their time to market and thus increase their profitability.  

HGST also has probably the highest cost to manufacture HDDs of any company now in the industry.  This is a result of their history of multiple sites in many countries making the components for the drives and then shipping these parts all over the world for assembly.  The company has made good strides towards consolidating its production facilities and this is starting to help the bottom line but this work is not yet finished. 
HGST needs to get its costs under control to improve their profitability.  

The obstacles to improved, much less leading profitability for HGST are daunting but Hitachi has taken steps to deal with these issues.  It will take continued determination to continue profitable quarters, much less achieve industry leading profitability. 

Everyone needs a dream and dreams can lead to fulfillment, especially if diligently carried out by a company with deep pockets.  We will have to keep our eye on Hitachi to see if they can turn their dream into reality.

Implications: * Iomega had previously entered into an agreement with ExcelStor Great Wall Technology owned by the Chinese government  * Iomega terminated this purchase agreement with ExcelStor  * Iomega entered into an agreement to merge with EMC  * EMC will probably use Iomega as a means to move into the small business and perhaps even the home storage market

Analysis: Personal storage and storage from small businesses is now dominating the growth in new digital content.  Since there are a great many more small businesses and individuals than larger corporations and since the costs of storage is decreasing and the need for storing increasing amounts of personal and small business content increases this trend will grow even more in the future. 

As a consequence of this growth in storage demand for small business and homes the market for external storage products to serve this market will grow as well.   As "Digital Storage in Consumer Electronics 2008", Coughlin Associates (www.tomcoughlin.com) projects over 60 M small external storage devices are likely to ship annually by 2010 with a total shipped storage capacity of about 42 Exabytes.  Many of these devices will be used to back up or aggregate content on home and small business desktop or laptop computers. 

EMC has experience in this area and has offered versions of its Retrospect backup software to the small storage device market for years.  It makes sense that the next move would be to invest in a company making external storage products.  

Although only a shadow of the former Iomega, inventor of the Zip and Jaz cartridges and drives, the company has a well known brand and offers some innovative devices based upon external storage.  These products include direct attached storage as well as network attached storage devices. 

Recent Iomega introductions have included media server boxes and other more innovative home storage based technologies.   EMC’s acquisition may be initially targeting small business storage but I believe over the next few years we will see more large storage systems companies making moves to establish themselves in the small business and home storage systems market. 

This will be a major growth area in the future and there will be tie-ins between storage in a small office and home and potential backup to larger storage systems on-line.  This will make the small storage system market more competitive and it may result in more strategic acquisitions of small storage system companies by larger ones.   Once again small is getting bigger!

Implications: * Hitachi introduced a 500 GB 2.5-inch hard disk drive in January 2008 but the product would not fit into a conventional 2.5-inch hard drive bay  * Fujitsu followed suit with a 500 GB 2.5-inch HDD announcement with a similar thicker drive package  * Samsung announced in March that they would be shipping a 500 GB 2.5-inch drive with the conventional drive thickness found in most notebook computers  * The Samsung product using 3-167 GB disks to achieve this storage capacity and form factor  * The suggested retail price for Samsung’s 500 GB Spinpoint M6 is $299 (about $0.60 per GB)

Analysis: 2.5-inch disk drives are the most popular drive format for lap-top computers.  This is also one of the fastest growing markets for hard disk drives as laptop computers displace “desktop” computers.  As a consequence of this growth in mobile devices it is expected that after 2010 sub-3.5-inch hard disk drives (especially 2.5-inch drives) will be the majority.  

Because of the reduced radius of smaller form factor disks the linear velocity at the outer radius of the disks is less than for a 3.5-inch disk.  This makes it easier to introduce higher linear densities on smaller form factor disks.  In addition higher track densities are possible since the flutter is lower for smaller disk form factors.  Thus new areal density increasing technologies are easier to introduce on smaller form factor drives.    The consequence has been rapid growth in 2.5-inch and smaller hard disk drive capacity. 

Today one can get 500 GB 2.5-inch drives and 160 GB 1.8-inch drives.  This is a 56% growth in available 2.5-inch storage capacity from 2007 and if this rate of growth is extrapolated a couple of years to 2010 the industry could deliver over 1 TB 2.5-inch hard disk drives.   Such drives will probably sell initially for $299 similar to the suggested price of the Samsung 500 GB unit today giving less than $0.30/GB for 2.5-inch drive storage. 

With the size of programs and content increasing and the increasing use of laptops to replace desktop computers this storage capacity will come in handy to many users.   Samsung has been a leader in introducing new HDD areal density technology over the last few years using technology from the remaining disk and head vendors (TDK for heads and SDK or Fuji Electric for disks).  At the same time Samsung has also been a major proponent of flash-based solid state storage, including solid state drives that are being promoted as competing against hard disk drives. 

By 2010 SSDs with 64 or 128 GB capacity could have prices similar to those of my hypothetical 1 TB (1,000 GB) laptop drive.  Clearly these storage products would appeal to different types of computer users!

Implications: * Seagate says that if SSDs start to threaten HDDs Seagate may sue SSD companies such as Intel and Samsung for patent infringement  * Seagate and WD have patents with many of the ways that a storage device communicates with a computer and SSDs may use some of these approaches  * SSDs are being introduced in laptops and other applications where hard disk drives have traditionally been used  * Seagate said that it would ship SSDs in 2008 and WD has had a long standing relationship with SanDisk, a major flash memory producer  * SanDisk has its own set of flash memory and possibly SSD patents through its M-Systems acquisition so there may be an interesting patent war brewing

Analysis: Seagate’s CEO, Bill Watkins, said that SSD companies may violate patents held by hard disk drive companies.  HDD companies such as Seagate and Western Digital have license sharing agreements on hard disk drive technology that allows them to use each others technology in many of their products.  Among the patent technologies referred to are communications protocols and methods for storage devices with computers.  Seagate has indicted that the company may sue SDD manufacturers if SSDs threaten HDD sales.  

Seagate and other hard disk drive companies were able to use patent violation suits to effectively stop the development of 1-inch disk drives by start-up company, Cornice a few years ago.  They may have in mind a similar approach to stop SSD companies that start to compete against HDDs. 

The situation with SSDs may be different than was the case with 1-inch hard disk drives.   Western Digital has a very close relationship and an extensive history with SanDisk, a major manufacturer and pioneer of flash memory devices, including SSDs.  Although WD has not announced a SSD product this relationship could give them access to SanDisk’s extensive patent portfolio and if WD creates a cross-license deal with SanDisk there could be an interesting foil to Seagate’s potential suit against other SSD makers.  

In addition SanDisk has its own extensive patented IP for flash memory cards and other devices and the company has a long history of going after and collecting from potential patent infringers.  An estimated 30% of SanDisk’s net profit comes from royalties for the company's many patents.  It is possible that Seagate could itself face potential patent suits from at least SanDisk when it rolls out its SSD products.  

The various players are trying to claim their IP and it is likely that there will be some suits pressed by many of the competitors in SSDs against each other as this market develops.  In the long run it is likely that SSD manufacturers will join in some sort of cross-licensing arrangement similar to that that HDDs makers have for many of their products.

Implications: *Western Digital introduced a 500 GB DVR expander product to add storage capacity to a digital video recorder with an external storage device  * Western digital plans to offer a 1 TB version of this product this year  * WD joins Seagate/Maxtor and others with external storage products to expand the storage capacity of DVRs and set top boxes  * These devices use the external serial ATA interface for a high speed external connection Keeping and organizing digital content is a growing challenge for consumers

Analysis: Set top boxes for cable systems more often than not now include digital video recording capability.  In addition companies such as TiVo make stand alone digital video recorders.  These devices are built to last for many years and only ship with enough digital storage capacity to hold tens of hours of programming.  Even the largest drives shipping in DVR products can only hold a few tens of hours of HDTV programming.  These products easily can fill up with recorded programming requiring the user to erase older material to make room for newer material.  During the years that these products are in use in a home hard disk drive technology continues to evolve, making higher capacity 3.5-inch hard disk drives available that can be used to record more programming.  Today you can buy a 1 TB HDD, but in 2 years 3 TB or higher drives should be available.  Replacing the HDD in a DVR is not something most users want to do and set top box manufacturers don’t encourage users to open their boxes.  Without some path to storage expansion users are stuck with limited program recording capacity.

Maxtor pioneered in the development of external expansion storage devices for digital video recorders and today Seagate, who acquired Maxtor produces these products.  Western Digital recently announced that they too are producing DVR expander products.  These products use an external SATA or eSATA interface with data rates of 3 gigabits per second to transfer data to and from the external drive to the DVR.  Most new DVRs are built with external SATA interfaces on their rear allowing these devices to be plugged in and providing expandable storage capacity for an older DVR.  With port multiplier technology multiple eSATA interface external storage boxes can be attached to a DVR allowing significant expansion of programming recording capacity over time.  Within a couple of years with 6 Gbps SATA and likely eSATA these external devices will provide very fast content transfer of many video streams at one time.

Cable operators and other service providers that buy set top boxes with DVR capability will benefit from external expansion storage devices in they can provide their customers with the minimal disk drive capacity within the set top box.  If customers need more storage they can just buy an external eSATA interface box and connect it to their DVR.  This reduces the cost of buying and implementing DVRs in the field.  This approach combined with the continued capacity growth of hard disk drives could result in  160 GB internal DVR drives with as much as 10 TB on external storage by 2014 according to Digital Storage in Consumer Electronics, Thomas Coughlin, Newnes Press, March 2008.

Capturing content on an external drive connected to a DVR provides temporary help to consumers that want to keep more content than they can hold on the internal drive but it doesn’t solve a larger problem in the home.  As we accumulate more personal and commercial content in the home; organizing, indexing and protecting this content becomes more crucial.  Moving into the second decade of the 21st century will require significant developments to make an integrated home storage utility a reality.

Implications: *  A start-up company, Pliant Technology announced a solid state flash drive for enterprise  applications  *  Pliant joins companies such as STEC in bringing flash memory to enterprise applications that require high throughput such as transaction processing  *  In January EMC announced that they would ship a version of their Symmetrix DMX storage system that could include STEC solid state drives as well as fibre channel or SCSI high performance hard disk drives  *  A new tier of enterprise storage is appearing that will first take on DRAM-based solid state drives but could eventually compete against high performance hard drives

Analysis:  First EMC announced that they would incorporate STEC solid state drives into their Symmetrix enterprise storage product line.  A start-up company, Pliant Technologies has announced $8M of funding to introduce their own line of solid state drives for the enterprise market.  Other companies have also introduced products for various enterprise segments.  One of the most popular is as local scratch memory in blades servers.  

The single level cell (SLC) solid state drives from STEC and now Pliant create a new level of the storage hierarchy for standard configuration Symmetrix storage devices and other heavy duty enterprise hardware.  These solid state drives use flash memory that performs writes and reads much faster than the flash memory used on USB drives or consumer electronic devices.  They do this by the use of special controllers in front of the flash memory chips that allow parallel data transfers to and from many flash memory lines at once.   

These products are not inexpensive and they are more expensive per bit than hard disk drives, even the more expensive fibre channel and SCSI hard disk drives, however they can provide higher data rates at a lower cost than is possible with even high performance hard drives.  Thus they are more economical than hard drives in $/IOPS (Input/Output operations per second).  

The new generation of flash-based solid state drives will compete initially against a small but important niche market of solid state drives made with Dynamic Random Access Memory (DRAM).  These are very fast in terms of operations per second but very expensive and large since they need backup batteries to maintain the data and often disk drives as well, and usually fit in a rack.  The new flash-based solid state drives are not quite as fast as their DRAM brethren but they take less space and are less expensive. 

A lower cost transaction intense memory device will open up this level of the storage hierarchy to more applications that couldn’t afford the DRAM solution.   In time if volume goes up and price goes down (including price per GB) flash based solid state drives may begin to threaten higher performance fibre channel and SCSI hard disk drives. 

It is possible that configurations with very inexpensive but much slower performance SATA disk drives could provide the mass storage behind fast transaction storage that consists of flash memory alone sometime in the next few years for some users.  This would be a sea change for the enterprise market.  Tom Coughlin, www.tomcoughlin.com

Implications: * Sandisk announced that they will start shipping 3-bit per storage cell NAND flash memory in March or April, 2008 * Currently the densest flash memory is 2-bit per cell multilevel cells (MLC) * The technology was co-invented with long-time Sandisk partner, Toshiba * The technology will be used on 16 Gb NAND flash chips using 56 nm photolithography and with 8 MBps write rate * Sandisk says that an advanced architecture for the storage devices allows them to achieve higher storage capacities without sacrificing performance or reliability

Analysis: Flash memory has been available for the last few years with single level cells having one bit of storage per cell (SLC) and 2-bits per cell multilevel cell (MLC).  The SLC memory is inherently more reliable but it is more expensive per bit of memory.  Improvements in the controller and memory architecture led to the ability to create multiple voltage levels per cells allowing more than one bit to be stored per flash memory cell.   

The write wear is worse as the number of levels in the cell are increased so the multilevel cell flash memory has to incorporate wear-leveling technology to try and wear the memory cells more evenly through the device.  The write performance is also worse for multi-level cells.  In order to provide faster write/read performance the flash memory must be organized to work in parallel so the net data is increased by the number of parallel lines of memory being used at one time. 

The wear leveling and parallel access are functions of the flash controller that stands between the host system and the flash memory cells.  Thus the key to advances in multi-level flash memory are more advanced flash controller technology.   If Sandisk can make good on its claim of equivalent reliability and performance with 3-bits per cell to 2-bits per cell then the company will have an advantage in the price per GB compared to flash memory companies that don’t have access to this technology. 

3-bit per cell technology requires no change in the basic design of flash memory from a manufacturing viewpoint, only in the architecture and controller side.  Thus the cost reductions per bit can be significant using 3-bits per cell (maybe 50% or greater cost reduction are possible).  With changes in manufacturing flash cell density alone (without the use of 3-bit cells) the cost per bit reduction may be about 30% annually so 3-bits per cell could lower the manufacturing costs for Toshiba and SanDisk significantly more than their competitors.  

SanDisk’s expertise in flash controllers was enhanced by their acquisition (completed last year) of M-Systems.  M-Systems had been working on multiple bit per cell technology before they merged with SanDisk.  It will be interesting to see what the response to this development is at other flash memory suppliers such as Samsung, Intel/Micro and (reportedly) Seagate. 

3-bit per cell flash will probably find its first home in consumer devices and USB storage devices.   Computer system manufacturers are still leery about using multi-level cell flash in enterprise and laptop applications.

Implications: * Encryption is used by many people and organizations to hide their information from unauthorized access  * Researchers at Princeton University were able to demonstrate a technique for finding the encryption key for software-based encryption programs  * With the encryption key these researchers have been able to access data that was protected by software-based encryption products  * The encryption key is accessed by chilling the DRAM on the computer to retain its memory so that an image of the contents can be created and the encryption key discovered

Analysis: Data encryption is a common way to protect data from unauthorized access.  When data is encrypted a mathematical formula is used to turn the data into a jumble of symbols that don’t mean anything.  The key to the encryption is used in this mathematical formula to decode the encrypted data to access the original content.  The more bits in the encryption key the harder it is for so-called “brute force” calculation techniques to find the encryption key by entering all possible combinations of the bits.  An encryption key of 256 bits length with the proper mathematical formula is considered very secure since any computer now in existence would take several years to crack the encryption key using “brute force.”  

There are several products on the market that implement data encryption by running encryption software on the host system.  This software encrypts the data on the host processor and stores the encrypted data on a storage device.  Many companies make such products including BitLocker (comes with Windows Vista), FileVault (comes with MacOS X and dm-crypt (used with Linux).  There are also commercial software packages that can provide software-enabled data encryption.   

Such software-based encryption packages may protect against casual attempts to recover protected data but it has long been know that there are ways to break this type of data encryption with the right sort of tools.  A recent paper from researchers at Princeton University points out one way that the encryption key can be recovered directly from the DRAM chips that are in the host system.  When DRAM chips are powered down they lose their stored data.  Actually the data is lost over a period of seconds to minutes.  If the DRAM is cooled, such as by an aerosol spray, the data loss can be made even more gradual.  Using such approaches it is possible to read the full contents of DRAM memory by copying the contents into another computer before the data goes away.  

If the DRAM memory contains an encryption key then this key can be recovered using this technology.  Note that the software tools mentioned above can be combined with other technology to eliminate their vulnerability to this particular hack but such research shows that software-based encryption may be vulnerable to discovery and penetration.  

So-called hardware-based encryption is less vulnerable to access and penetration.  One approach to hardware encryption is promoted by the Trusted Computing Group and involves putting the encryption key on the storage device which also contains a special chip that does the encryption/decryption.  Such a device would not expose the key to the host system and so techniques such as reading the key from DRAM will not work.  Since the key can be stored on a non-user accessible area of the disk drive reading the key is nearly impossible.  Also if the key on the storage device is erased the data on the device is crypt-text which cannot be reconstructed as information without the “brute force” techniques discussed earlier.  

Hard disk drive companies are the first companies to introduce products with on-board data encryption.  The initial products from Seagate and Hitachi are 2.5-inch disk drives for the mobile computer market.  These products were targeted first because of the high incidence of theft of laptop computers.  Encryption on the hard drive means that the data in these hard disks cannot be accessed if the laptop falls into the wrong hands.   

Although initial storage products with on-board encryption will be hard disk drives the Trusted Computing Group is also creating standards for tape, flash and even optical disks with on-board hardware encryption.  Such products could do much to increase the level of security of computer users.

Implications: *The HD DVD Format, promoted by Toshiba and several other consumer electronics companies has been competing with the Blu-ray format for the next generation of HD video distribution  *Just before the CES in early January Warner announced that they were pulling their titles from the HD DVD format by May 2007  *Following the Warner announcement other studios have indicated they were considering switching formats and recently Netflix announced that they would not carry HD DVD titles  *Toshiba announced that they were halting HD DVD production and then said that they would abandon the format  *This ends a two year old format war for the next generation of high definition optical disk distribution

Analysis: For the last two years Sony and their partners have been pushing the Blu-ray high definition video distribution format while Toshiba and their partners have been competing with an alternative HD DVD format.  During this time each format has obtained support from various studios and distribution sources.  This has led to a common sight on retailer shelves of a small collection of two more expensive high definition disk format movies.  The two formats are very different and except for a offering of more expensive players by LG and others that can play both formats the formats are incompatible.  Players for either format could also play the older DVD format discs.

Sony produced a product that is more complex and at least initially more expensive to produce.  Blu-ray disks offer a path to higher storage capacity than HD DVD disks and thus the technology could be expanded for more or richer content. They installed Blu-ray players in their PS-3 game machines that could also be used to play movies on HD TVs.  The use of Blu-ray players in PS-3 gave Sony an early lead over HD DVD in unit sales, including disc titles.  

Until the end of 2007 it still appeared that the format war would continue for several more years.  A format war discourages consumer purchases so neither format can go up the volume curve to lower production costs and thus lower prices for consumers.  Consumers are very price sensitive in their purchases of such products and they don’t want to buy both formats, thus the market has been stuck in a slow growth mode.

The announcement by Warner that they were dropping titles from the HD DVD format dealt a severe blow to Toshiba and the rest of the HD DVD partners.  Apparently many of the other content owners have been examining their options with the possibility of a mass migration to the Blu-ray format.  Recently Netflix announced that they wouldn’t carry HD DVD titles any more.  

Toshiba sensed that the writing was on the wall and just announced that they were halting production of HD DVD drives and within a day they said that they were abandoning the platform. This retreat by Toshiba left current customers feeling bad about their purchase decision since their products will rapidly become obsolete as new titles will not appear in HD DVD.  These customers will be looking to Toshiba to offer them some help in their transition to Blu-ray.  It is not clear what or if Toshiba and their partners will help these customers but it would be beneficial to their long term reputation if they made some efforts to do so.  However, the existing customer base for HD DVD is still relatively small and the selection of a single format is beneficial to the growth of the industry. 

Internet downloading and streaming of content, even high definition content is growing and the optical disk distribution market must work hard to maintain their competitiveness.  A 25 GB file may not be a big deal to download through the Internet or a Kiosk in the next few years and optical disk designers must develop the next generation of high resolution content and distribution media to stay ahead of electronic distribution.  Blu-ray gives them a path to higher capacity but eventually it too will run out of gas.  The next generation HD specifications are already being developed in Japan and they may need distribution media that can hold hundreds of GB or even a terabyte of data to stay competive.  The war between electronic and physical distribution continues.

Implications: * Yoichiro Tanaka, senior manager of Toshiba’s Storage Device Division gave a talk at an IDEMA Japan event  * He disagreed with three analysts that spoke before him that HDDs were threatened by flash memory  * While Tanaka-san recognized the important role that flash memory provided for mobile devices for playing content he pointed out the result of these flash based players was a net increase in the amount of information that is distributed and stored  * He stated that HDDs will contain 80% of the world’s information * The prediction is based upon increasing adoption of 1.8-inch hard disk drives in camcorders, notebook PCs and external HDDs—he stated that in 2008 more than 40% of all camcorders will have embedded HDDs.

Analysis:  For the last couple of years there has been a pretty steady barrage of comments and analysis suggested that HDDs are soon to be replaced by flash memory in all sorts of consumer as well as computer applications.  The argument is that flash memory is increasing in storage capacity and decreasing in price, making it affordable for many applications.  In this view of the world hard disk drives will be gradually displaced by solid state memory in application after application.  

In actual fact HDDs have been exhibiting steady unit growth and also growth in storage capacity.  In 2007 I estimate that there were about 508 million drives produced.  This is 16% more disk drives than were shipped in 2006.  At the same time by the end of 2007 hard disk drives with different form factors offered 1 TB (3.5-inch), 500 GB (2.5-inch), 160 Gb (1.8-inch) and 40 GB (1.3-inch). 

I project that with rising demand for hard disk drives to hold the world’s mass storage and to support applications using flash memory such as music and video players as well as digital still camera there will be over 1 billion drives shipped annually by 2012.  By then there will be several TB capacity 3.5-inch drives, more than 1 TB in a 2.5-inch drive.  1.8-inch drives will offer hundreds of GB and there could be more than 100 GB on a 1.3-inch drive.  

In my view and it appears to be shared by Tanaka-san there is much more synergy in the market between flash and hard disk drives.  The growth of either makes the other more valuable and helps drive demand for it as well. Both hard disk drives and flash memory offer increasingly more cost effective storage media that will find more and more applications.  However most flash applications will require the mass storage capability of a hard disk drive to provide content or to save recorded content captured on flash.  Likewise inexpensive hard disk drives enable consumers and companies to provide high resolution content that can be stored for playback on mobile devices using flash memory.   

Furthermore, as is the case for Turbo memory, Hybrid Hard Drives, SanDisk’s Vaulter and other products there may be overall system advantages to combine both types of storage together to provide a higher performance and higher capacity total product solution.  The market for digital storage is not so simple as “winner takes all”.  Flash and hard disk drives are increasing in storage capacity at about the same rate so they may never reach parity. 

At the same time increasingly less expensive storage encourages product designers as well as content creators to provide ever richer and higher resolution experiences.  I see no end to this innovation anytime soon and inexpensive storage is one of the key enablers.   Thus I applaud Tanaka-san and Toshiba for defending the valuable role of HDDs and particularly defending small form factor hard disk drives. 

It is important that hard disk drive companies not let the 1.8-inch form factor be taken over by other memory applications for as history has shown smaller form factor hard disk drives are the future of this storage technology due to the physics of disk drive design and scaling.

Implications: * Hitachi announced a 2.5-inch form factor disk drive with single drive capacity of 500 GB—the highest so far in this form factor  * Hitachi added another disk (3 total) and heads to achieve this with a resulting drive thickness that is 3 mm thicker than the industry standard 2.5-inch for laptops * The drive is expected to initially sell for about $400 and includes bulk data encryption and is rated for higher usage conditions * Asus, a Taiwanese laptop manufacturer is offering a dual-drive laptop with 1 TB storage capacity * Hitachi also offers a 400 GB version of the drive

Analysis:  Hitachi Global Storage Technologies led the pack in the introduction of a Terabyte 3.5-inch hard disk drive in 2007.  In January 2008 they announced a record-breaking 500 GB 2.5-inch hard disk drive.  The prior record for 2.5-inch hard drives by Fujitsu and then Toshiba was 320 GB.  Shortly after the HGST announcement Samsung said that they would also introduce a 500 GB 2.5-inch drive.

Asus, a Taiwanese laptop manufacturer announced they would deliver a TB laptop using two of the new Hitachi drives.  In addition to traditional uses for 2.5-inch drives in notebooks such significant storage capacities in a small form factor makes smaller desktop computers possible as well as smaller consumer devices such as DVRs and set top boxes.  These products could also find application in smaller form factor ATA drive arrays in enterprise applications and in external mobile storage devices.

I expect that other HDD manufacturers including Toshiba, Fujitsu, Seagate and eventually WD will be introducing their own 500 GB 2.5-inch drives within the next few months.  There are also rumors of a 1.5 TB 3.5-inch drive and even a 2 TB 3.5-inch drive being introduced by one of the hard disk companies sometime this year. 

The disk drive areal density growth continues leading new product introductions at a rate of about 50% annually, especially for smaller form factor hard disk drives. Storage capacity increases, especially if done without adding disks and heads is an important way to create product differentiation and thus increase market share. 

Hitachi recently announced that they were dropping 1.8-inch disk drives but the new announcement shows that the company continues to develop advanced hard disk drives in the more focused and higher volume 2.5-inch and 3.5-inch HDD form factors.