Everyone has no doubt gotten the word about today’s announcement from Isilon about their next-gen clustered NAS product line. If not, here are their talking points in a nutshell.
I. Isilon Continues to Deliver on its Long-Term Vision
a. Achieved profitability in Q4 ‘09
b. Record revenue and channel contribution in ‘09
c. Two new product lines released in March ’09 accounted for 50% of ’09 revenue
d. Nearly 1200 customers worldwide, with more than 260 new customers in ‘09
II. Isilon Capitalizes on Market Demand
a. Two out of every three customers in ’09 were non-Media & Entertainment
b. Rapid adoption in server virtualization, life sciences and government
c. North American channel grew from 25% of revenue in ‘08 to more than 55% in ‘09
III. Isilon Continues Legacy of Innovation
a. Introduces new products delivering SSDs for scale-out NAS
b. Leverages SSDs in combination with either SAS or SATA drives, balancing performance and cost
c. Continued innovation keeps Isilon’s scale-out NAS solutions generations ahead of major incumbent competition
Now, before you shrug this off as just so much more marketecture, let me tell you why I am following this company. (By the way, I have no financial relationship with them.)
First, everyone I am talking to right now seems to be interested in clustered NAS. Files are filling the junk drawer at a much faster rate than block/transaction data, making NAS a desirable storage solution. However, hanging out a lot of NAS mounts with no global namespace and no common management method is a pain. Hence, the world is looking for a good cluster/namespace story — of which there are precious few.
The cloud guys know this and are doing their best to beat NAS with NAS on steroids. Not sure how that will all work out, but I do know that the technology to actually cluster and namespace has been dragging its butt for years.
Isilon seems to have cracked the code. According to big customers, Microsoft, ARM, etc., they are delivering the bang expected from the buck. IBM just announced a competitor in SONAS, which builds on a less than successful technology called SOFAS (sp?). Snapshot-based backup via Tivoli is built in. We’ll see how that goes. Note that GX clustering and new/forthcoming ONTAP features in NetApp land are also supposed to compete in this space. However, if Isilon is to be believed, they are winning deals against NetApp after customers do some comparison testing.
Also, Isilon’s (and IBM’s) news has a kind of weird tie-in with developments with Dell and Exanet. Recall that Exanet burned through $70M of VC investment and went belly up. Dell is now picking over the carcass for less than what Exanet was asking them to invest to keep their company out of Chapter 7. Exanet was also trying to press forward a NAS clustering/global namespace play.
The other thing I like about Isilon is their use of Flash SSD. They are using it simply to store metadata about files parked in their clusters. That is not a hugely transactional use case, and it seems to me to be the first intelligent use of the technology that I’ve seen. Others have been proffering Flash as a database accelerator, but as I have argued here before, Flash SSD’s memory wear problems make it less appropriate for lots of changing data. I go for DRAM SSD for database acceleration, not Flash, otherwise I will probably be changing out the Flash drives every couple of hours/days/weeks depending on my workload. Kudos to Isilon for the first real-world application of Flash SSD that I have seen.
Still, it’s a shame that Isilon doesn’t offer their software as a standalone product. That way, I could use whatever storage I wanted.