Seagate, Fujitsu discuss SSD strategies

Tag : drives & storage devices

Rich Vignes, senior manager of market development in SeagateTechnology's solid state group, says the company looked at SSDopportunities for at least 18 months with the initial thinking thatthe technology would displace its HDD business. But instead, thecompany came to the conclusion that it will be additive to, and infact expand the overall storage market. "SSDs will create a newtier of performance," he told SST . "They will be utilized in several product variants in enterprisearchitectures and SSD early adoption will leverage existingecosystems."
Seagate believes that in the next few years SSD adoption will comein the form of a storage device that has the same interface andform factor as hard drives, making them easy to plug into theexisting ecosystem without any change whatsoever. "Then, as peoplestart to optimize architectures, that picture will likely change,"noted Vignes.
With respect to its own product portfolio, Seagate sees a place forSSDs only in tier 0 enterprise (i.e., mission critical) data centeroperations ( see figure below ), particularly for performance requirements in the range of 5K-15KIOPS. "This is where we believe SSD has its best value proposition-- with outstanding performance, reducing system level power, andreducing the total system data center footprint," Vignes told SST , adding that the other segment tiers are capacity driven andtherefore better suited to HDD storage. However, the companybelieves that the growth of SSD share in tier 0 will also enablethe growth of HDD storage in tiers 1, 2, and 3 because tier 0 datahas to be backed up onto these other tiers. "So SSDs play a verysmall, but very important role in the total storage stack."


Figure 1. SSD and the Seagate portfolio. (Source: Seagate)
CLICK HERE to view larger image

Seagate's portfolio strategy reflects its market strategy. "Ifyou're a new SSD provider, who previously hasn't sold storage intothe enterprise, this is a new market and a new product for you,"Vignes told SST . "Whereas for Seagate, it's only a new product, not a new market."The company plans to serve the same end-user base, and believesthere are many similarities between an SSD and an HDD so it canleverage its expertise and IP and direct them to buildenterprise-ready SSD storage. "We absolutely believe that SSDs andHDDs will co-exist for the long-term," he said. "We continue toimprove the areal density of HDDs, and while SSDs do experience adrop in their cost because of improvements to NAND tech, thosecurves stay parallel for as far as the eye can see."
Fujitsu: SSD "advantages" illusory for consumers
For Fujitsu, which has a development team investigating the marketpotential for SSD, the upshot is that there will be two marketsegments that will gravitate to SSD: 1) thin, light, low-costnotebooks; and 2) enterprise-specific applications.
For now, however, "Fujitsu doesn't think there is much opportunityfor SSDs in the notebook space," according to Joel Hagberg, VP ofmarketing & business development at Fujitsu. The company believesthat SSD penetration won't be a factor in the notebook space until2012, and thinks there is a lot of hype regarding the replacementof the traditional HDD in the mobile space with SSD. "There's not alot of opportunity from Fujitsu's perspective," he told SST . "Depending on which analyst's data you use, there have only beenabout 120M+notebook drives shipped last year and of those,~98,000-140,000 were SSDs, a very small percentage of the total."
Hagberg noted that at this year's Flash Summit there was debateabout what really constitutes an SSD. If the answer is "a flashchip on a motherboard," then that will quickly influence the volumenext year, he observed. He also noted that cell phones andultra-mobile PCs aren't large enough to hold an HDD, so theseweren't a market for HDDs anyway.
Consumers are having a hard time spending the money for perceivedadvantages in the notebook space, Hagberg told SST . He specifically cited performance advantages commonly associatedwith SSDs that he says are not panning out for consumers -- forexample, consumers don't really see faster boot times for theirnotebooks (and in fact, boot times might even be a little slower),and the performance play in traditional applications such asPowerPoint or Word also aren't there. Moreover, Hagberg said,because HDDs go into sleep mode and draw 1/10 of the power when inthe read/write state, even the "advantage" of SSDs in terms ofpower and battery life are illusory. "All power measurements ofHDDs vs. SSDs are done during read/write," he explained. "SSDs haveone running state: a steady-state draw of energy, so HDDs have anextended battery life as a result."
In the enterprise space, Fujitsu sees two big growth areas: 1) 2.5"10,000rpm mainstream server drives, and 2) the entry level SATAdrive. By 2012, the industry will see some critical mass for SSDfor high IOPS. "SSD works very well when you are reading [smallblock reads with very high IOPS], but not writing data," saidHagberg. "HDDs work very well for large file transfers and also forwriting data." Citing some analysts' observations, he noted thereare doubts that the NAND industry can afford to take 10% of the HDDenterprise market share, given the fab capacity that would beneeded by 2011 to achieve that.
In summary, Fujitsu believes HDDs will dominate volume in the nextfew years and new technology will be needed to overcome concernsabout SSD reliability, cost, and performance. The company is alsoevaluating alternatives to NAND flash. -- D.V.