Solid State Drives (SSDs) are not really news to those of us on this side of end-products, but some medium-term market strategies in the data storage sector may raise an eyebrow. As we continue to watch the U.S. slowdown, global economic volatility, soaring oil prices, currency devaluation, and consumer spending negatively affect the semiconductor industry through depressed profitability and jittery forecasts, a few major players are actually whistling a happy tune. Those who are poised through well-planned, medium-range strategies that take advantage of recent price collapses in memory, new business alliances, and that have positioned new products are likely to ensure themselves leadership positions into the next decade.
The battle for data storage, particularly in mobile media and notebooks, is not only here: it is about to flare. According to Tech-On, Toshiba’s President Shozo Saito is predicting 25 percent of "all notebooks shipped to be equipped with SSD by 2011" (http://techon.nikkeibp.co.jp/english/NEWS_EN/20080421/150769/). Lehman Brothers forecasts SSDs to "replace 2% of [Hard Disk Drives] HDDs in the computing segment in 2009, 4% in 2010, and 10% in 2011" (Lehman Brothers HDD Components Industry Outlook 1/15/2008, p.1). Regardless of how conservative or eager your estimates are, the "post-HDD era" has already been coined – but that doesn’t mean that HDDs are about to disappear.
Data storage market overview
HDD has been a tried and true technology for data storage in computing for many years, and will likely continue to hold onto a decent market share for the foreseeable future. HDD is a mature technology, and therefore reaps the advantages of consumer trust and profit margins; the HDD sector has experienced double digit volume growth for eight years and is projected to continue at this pace through 2009, according to Lehman Brothers (LB, ibid., p.1). However, this level of growth is now unsustainable and is expected to slow from over 12 percent "to 8.3% in 2010 and 6.3% in 2011," due to SSDs replacing HDDs in computing (LB, ibid., p.1). One important industrial consequence of this product shift is the likely further reorganization and consolidation of the HDD industry to "3-4 companies." (LB, ibid., p.16)
Maturity does not exclude innovation though, and HDD technology has continued to innovate. Most notable are the recent increases in surface-recording density based on new head technologies: tunneling magneto resistance heads (TMR) and perpendicular magnetic recording (PMR). The expected rapid adoption of these new head technologies will improve HDD's competitive position and marketability in the face of SSD adoption. While HDDs face a physical limit at 6TB/square inch (7-10TB 3.5-inch HDDs), this leaves room for growth driven by steady improvements in surface-recording density over the next ten years (LB, ibid., p.12).
Apple iPod's switch to NAND flash memory in 2005 marked the real transition from HDD to SSD. As Apple eliminated HDDs in all but the iPod Classics, other mobile media followed. Apple not only revolutionized the personal music player space, but with the success of the iPod line, NAND flash memory also benefited through increased awareness and product trust. As a result of this shift, the maximum utilization for 1.8-inch HDD is believed to have been reached in 2006 (LB, ibid., p.13). Evidence supporting this position can also be found in examining companies such as Hitachi GST, which "shed it's [sic] 1-inch product line late last year" (www.eetasia.com/ARTP_8800503993_499486.HTM).
Growth is expected to continue in the 3.5-inch and 2.5-inch HDD segments, with leaders like Hitachi GST positioning themselves to respond to strong demands. According to an interview with EE Times-Asia and Robert Chu, VP of Asia Pacific Hitachi GST, the consumer demand for staple products such as "DVRs, camcorders, STBs, multimedia systems, home servers and external media storage [as well as new consumers in China, India and emerging markets of Southeast Asia] will continue to propel the growth" (www.eetasia.com/ARTP_8800503993_499486.HTM).
While SSDs have reached the consumer, business and industrial markets, some barriers remain. For the industrial consumer, particularly in the military and aerospace sectors, SSDs are not new technology at all, nor is price a determining variable in the decision making. Defense and aerospace sectors have long used SSDs for mission critical purposes because of the vibration, impact, and heat tolerance levels coupled with the high levels of data integrity and preservation (http://i.cmpnet.com/informationweekreports/doc/2007/159SSDTT.pdf).
The business consumer is more willing to take on some higher price points and risk-reward opportunities because of improvements in data integrity/data loss avoidance, power consumption, speed, durability and performance. Two examples of the inroads by NAND flash memory for data storage for the business consumer can be seen in adoption of USB flash drives, and to some extent the success of blade servers, at the individual and firm levels, respectively.
The advantages for the consumer, particularly for notebooks and mobile media, are obvious even to the novice buyer of consumer electronics. Regardless, price point coupled with the uncertainty of a new technology is a real force in determining consumer purchasing behavior, particularly in today's spending conscious economy. High prices aren’t the only place to lay blame for the fact that market penetration remains slow: the presence of other challenges for SSDs, including real performance and speed differences compared to cost, and the present limitations on overwrite capabilities (an estimated five-year lifetime for the average notebook – a tough sell to many PC makers facing mounting challenges in sales) are hindrances to SSDs adoption.
The challenges ahead for data storage
iSuppli's 1Q08 market outlook showed a 7.4 percent decline in global DRAM revenue, with no DRAM supplier generating a profit. "If the market hasn't bottomed out yet, OEM DRAM buyers will encounter major risks of DRAM supply chain disruptions because some suppliers are facing serious financial issues, and potentially could leave the business" (http://www.isuppli.com/news/default.asp?id=8933&m=5&y=2008). Despite the continued DRAM troubles, some firms, such as Samsung, Elpida, Hynix and Powerchip are making strategic moves to corner market share. Others, such as Semiconductor Manufacturing Corporation, have announced that they are exiting the DRAM market.
DRAM is not forecasted to reach pre-2007 average selling price (ASP) levels, and that has a significant impact on the data storage sector. With the shake-out of the DRAM supplier market, some have shifted their production to NAND flash memory. The NAND shift has contributed to the continued ASP dampening. On the other hand, the increased focus on NAND production has also promoted R&D in both NAND technology and production improvements. As NAND nanotechnology advances, and increased investments are made in the shift from DRAM to NAND production, SSDs stand to gain from the increased memory capacity and lowered costs and ASPs.
Of course, the past three quarters have also seen significant NAND price declines; however, lead firms in particular are using the ASP drops as opportunities to pass savings onto the consumer in the hopes of boosting adoption through lower price-barriers and/or increased memory capacity offerings. Intel has certainly aligned its strategic plans to turn possible-recession, inventory gluts and NAND ASP declines into market penetration opportunities for SSDs in 2008. According to CEO Paul Otellini, Intel plans to offer a new line, called High Performance SSD, which is intended primarily for mid- and high-end notebooks and enterprise storage. By the end of 3Q08, these SSDs, with increased capacities and speeds (http://arstechnica.com/news.ars/post/20080310-analysis-intel-to-fight-recession-with-solid-state-drives.html), will be bundled with Montevina, the Centrino 2 platform, according to DigiTimes on 23 May 2008. According to the report, the Montevina bundled SSDs will be based on a SATA interface, offer an 80GB capacity and come in two sizes, 2.5-inch (Client X25-M) and a 1.8-inch (Client X18-M). Also announced in the DigiTimes article, "Intel plans to increase storage capacities up to 160GB by the end of the fourth quarter, and to 250GB and above in 2009." Other sources have identified other new SSDs from Intel that will be released in 2H08 with significantly higher speeds and based on the Open NAND Flash Interface 2.0 (ONFI 2.0) "which will offer 200MBps of sustained read throughput and 100MBps of write throughput" (http://arstechnica.com/news.ars/post/20080201-new-high-speed-flash-gets-giant-speed-boost.html). Intel is, however, not alone in the new SSD offerings set for 3Q08, Samsung announced on 25 May 2008, the upcoming release of their new Multi-Level Cell (MLC) 1.8-inch and 2.5-inch SSDs equipped with SATA II interface (www.tgdaily.com/content/view/37619/135/). These new drives are now the fastest performance SSDs, with read rates of 200MB per second and "sequential read rates of 160 MB/s" (ibid.). Couple these technology advances with lower NAND pricing, and two significant barriers to SSD adoption are lowered considerably; it should be noted that SSDs are still ranked the highest in price.
There are, however, other technological developments that must mature. In order for SSD to gain the momentum projected and take the market share that analysts portend, the shift from Single-Level Cell (SLC) to MLC technology must be completed. However, MLC is not likely to be the final answer even though MLC allows for higher capacities than SLC chips. MLC brings with it "much higher bit-error rates, however, and can only be written about 10,000 times – one-tenth the number of cycles for single-level cell flash" (http://www.automotivedesignline.com/204701376). Many analysts agree that a more immediate solution will likely be found in SLC/MLC hybrid chips, allowing time for the momentum to build and bring MLC-based SSDs to the mainstream (Lehman Brothers, ibid., p.14). Furthermore, SSD will continue to benefit as increased nanotechnology advances, at the core of NAND flash memory research, provide yet unforeseen solutions.
A final hurdle to SSD may come not from present knowledge, architecture or technological limits, but rather from a legal dispute. Seagate Technology filed a patent-infringement suit against STEC, Inc. (formerly SimpleTech, Inc.) "alleging violations of patents covering storage devices" (http://online.wsj.com/article/SB120821108792914215.html). Seagate Technology, which acquired Maxtor in 2006, and Western Digital are both leaders in the data storage sector and both hold numerous patents that directly relate to SSD technologies. The challenge to SSD advancement now lies in the ability and willingness of firms to collaborate and to honor patent-protected innovation. The question for the lawyers is where the line is to be drawn and what constitutes fair recognition and compensation.
Making the move
The move from HDD to SSD is not just a simple product shift, it has important reverberating effects throughout the manufacturing, components, memory, and R&D sectors of the semiconductor industry. Not only does the move away from HDD involve a significant number of players (from DRAM memory chips to spring, head and related components manufacturers), it also further opens the market reorganization in memory already begun by the ASP plummets of DRAM and NAND and the high-cost of nanometer technology (cf. MarketWatch Quarterly v.1 nos. 2 & 3).
Presently, the market is still relatively segregated between SSDs and HDDs, though the inroads are likely to deepen even over the next few quarters. At the end of 2007, Gartner had projected "the tipping point for SSDs at 2009, though other analysts do not expect widespread adoption until 2011 or 2012" (www.automotivedesignline.com/204701376). One of the main reasons for the delayed penetration is the price per gigabyte for SSDs as opposed to HDDs (ibid.). HDDs are likely to hold strong in the desktop sector well into the next decade, based on price/gigabyte, product maturity and product requirements that are well met by HDD. Notebooks and the mobile sector, however, are the battlegrounds – though likely not for long with SSDs projected to already make a 5 percent inroads into "the overall notebook market" by the end of 2008, according to American Technology Research (as printed in www.automotivedesignline.com/204701376).
While investment in 1-inch and smaller hard drives for HDDs has mostly been curtailed by the larger firms, the HDD industry continues to perform well. "Momentum has accelerated in 3Q and 4Q 2007 after production adjustments in 2Q" (LB, p.1). While the numbers may be strong with continued double digit growth for 2008 and 2009, the handwriting is on the wall for HDD, and manufacturers throughout the HDD supply chain have already implemented medium-term strategies that allow for core practices to be diversified to related industries and products for the 'post-HDD era.' Lead firms are certainly already jockeying for best supply chain, pricing, and product strategies during the next three to five years. Opportunities for new market share and leadership positions are quite promising for those able to take advantage of the present economic market volatility. For others, significant reorganizations in DRAM and HDD component sectors could signal additional firms' exiting the