Nimble Storage announced their All Flash Array in February, the AFA3000/5000/7000/9000 series, and the Nimble Storage vs. Pure Storage fight is on.
Pure Storage has been out in front of the all flash array (AFA) market for some time now, a position that has energized EMC and pressured all the competitors to match price or features or a combination of both. The field has been crowded, but many have fallen to the back of the pack, and to continue the NASCAR analogy, many will spin out, crash, run out of fuel or just plain finish last. XtremIO from EMC has come on stronger than any other vendor, and the FUD has been flying back and forth like trash talk at a prize fight. SolidFire was bought by NetApp, another step along a confusing product strategy path from the parent company. Pure arguably still owns the market, but questions have come up concerning profitability – is Pure winning the deals but losing money? Not according to them.
Nimble Storage has dominated the Hybrid Flash Array market since it hit the market and has had enviable success until just very recently, where EMC came out swinging and delivered some body blows with THEIR “win at all costs” strategy, trying undercut Nimble in a no-compromise strategy that fended off Nimble’s growth into the Enterprise market. Not my conclusion…the finance analysts’.
Nimble’s strength has been providing flash performance at a hard drive price, and that value has been proven over and over in the last five years. Great support, a great product and a great channel strategy has made them a terrific partner in the datacenter market and created a lot of happy customers. Nimble Storage has shown market leadership in their cloud based analytics and reporting solution, Infosight, which is free to all owners of Nimble Storage.
Nimble Storage is Scale Out and Scale Up, which isn’t offered by anyone else in this market. This means that you can add storage shelves to expand capacity, or buy another Controller pair to add processing capability to create one virtual array. (NetApp can do it, but it isn’t really the same kind of scaling and is expensive, complicated and has not had widespread adoption.)
Nimble’s other value points have been: all software included, easy to use, great support and user community. Their reliability is nearly six nines, unheard of except perhaps with the enterprise storage systems such as VMAX and Hitachi – 99.9997% - as reported in Infosight!
Pure’s main value to the market has been first-to-market with a viable, reliable and affordable (relatively) all flash array storage system. Pure Storage also has some common value points: fantastic performance, very easy to use, all software included, great support and user community. They, also, can boast very high reliability numbers.
Flash Gets Cheaper, So Now What? Nimble’s Response
As the price of flash has decreased, and the efficiency of data reduction software on the arrays has improved, the USABLE capacity of flash storage has increased to the point where the EFFECTIVE $$/GB is close to Hard Drive $$/GB. New SSD technology is on the market with 3.9 TB drives, and 15.9TB is hitting the streets. So, how does Nimble Storage respond to the fact that hybrid flash isn’t as important in this new age of cheaper, high density flash?
- Density – at least for now, is in Nimble’s favor. The top stack in the picture below is Nimble. Pure, XtremIO and SolidFire are below it. How?? First to market with 3D NAND flash drives, very high capacity flash technology. That advantage will likely diminish soon, but for now, they are ahead in the arms race. There are other technical efficiencies around memory utilization, but let’s leave it at that for now.
- Infosight – still nothing like it in the market – Pure has a customer portal, and a partner portal, but not quite as good or as complete as Infosight. That is going to change soon as well (rumors of mobile app to check the Pure Portal), but perceived advantage goes to Nimble in the early rounds.
- Scalability – Nimble still will scale out and up. Pure chose not to go down that road for some technical reasons that translate as increased risk for availability, but not sure if that is relevant or probable yet.
How Does Pure Storage Compare?
- Pure still has (arguably) better dedupe, the primary data reduction strategy. Nimble can set up “dedupe domains”, specific by volume/block size/application, but it remains to be seen if that is as effective or efficient in the wild (their numbers say it is).
- Non-disruptive upgrades (NDU) – Nimble can’t do “cross-generational” upgrades without downtime, due to some technical reasons. Nimble arrays can be upgraded all day long within the family non-disruptively, and without performance impact, but going from old to new (e.g. CS400 to CS500) would likely involve downtime. Pure has built in upgrade or downgrade NDU.
- Enterprise track record – Pure can point to VMAX or NetApp and other big iron replacements. EMC worked hard to block Nimble in enterprise accounts in the last two quarters, so they don’t have the track record (yet) that Pure does.
- Pure just released the new FlashBlade – 130 PB in a rack. Very new, but it might have a huge impact. This new product is positioned as an Isilon (and NetApp) killer – coming to market at a $1/GB price, NFS/CIFS protocols. Scaling and block connectivity are coming very soon. This should be in directed availability soon, and GA this summer.
Nimble Storage vs. Pure Storage: Which is Faster?
That depends. Nimble does fantastic on writes, and will now do great on reads with all flash. It remains to be seen which is faster, but both are plenty fast enough.
Nimble Storage vs. Pure Storage: What about Price?
That also depends on the model. Pure just released the //m10, a roughly $50-60K array (street price). The Nimble AFA3000 is in the same range for price and capacity.
If you have questions around the pros and cons of Nimble Storage vs. Pure Storage (or any of the other players), leave a comment or reach out - I'd love to hear from you!