First, some links for background as to what’s going on in the industry:
- EMC trips and kicks their supplier, STEC
- STEC shares crushed as EMC carries over orders into 2010
- Samsung Makes Strategic Investment in Fusion-io
- SSD Buyers Guide on StorageSearch.com
Solid-state technology is far superior to traditional hard drives in random access time, sequential and random writes and reads, and vibration resistance. As of today, there are still two major factors that are holding solid-state back from widespread adoption, cost and reliability.
Hard Disk Drives (HDDs) have been the mainstays of computer-based long-term storage since the personal computer was invented. They have followed a similar path to Moore’s Law on the CPU side, where roughly every 18-24 months, hard drive capacity would double, which would in turn cause prices to collapse. This was driven by an insatiable demand by consumers and enterprises alike to be able to store more and more data. But how long is that sustainable? When do customers start to say “I don’t need any more storage”? Now, I fully realize that Bill Gates also said “640k ought to be enough for anybody” back in 1981, and we see the obvious fallacy there. However, I think the main point is not how much storage each individual person needs, but where do they need it?
I believe that consumers and enterprises will both continue to need increased storage (storing endless amounts of pictures, videos, music, etc) but I think that users will start to migrate towards a more “online” approach.
Thus, there is an intersection for solid-state drives where consumers and enterprises have “enough” space for short-term storage, and then they will look to online, or archival (probably through very large hard drives) methods for long-term storage. We haven’t hit the point where SSD technology is cheap enough for this to really make sense.
On the consumer front, I believe that most users will be able to get by on a 128GB SSD. When this drive hits $150, it will gain widespread (>20%) market share within two quarters.
On the enterprise front, I think reliability is still the key issue, as cost to performance ratio is already at parity with equivalent SAS/FC solutions. I recently visited Oracle OpenWorld up in San Francisco, and every major vendor had an offering utilizing solid-state technology. At this point, it’s still new technology, wear-leveling, crash tolerance and a method of measuring risk are still brand new.
2010 will be the true rise of solid-state in the industry, with 2011 seeing most desktop, notebook and enterprise systems incorporating solid-state in some way. The hard drive manufacturers already know this, note WD’s purchase of SiliconSystems and Seagate’s plans around Solid-State.
Disclaimer: I already have a 128GB Crucial SSD in my Windows 7 notebook and it SCREAMS