Forget software, flash storage is eating the data center (which is software in its own way).

The winners in this market appear to be the AFA players:

• Pure Storage supposedly working on its S-1
• XtremIO nearing $1B in annual revenue
• SolidFire endorsed by more funding

On the ropes:

• NetApp’s FlashRay was folded into ONTAP (and project lead Brian Pawlowski defected to Pure)
• Cisco’s Whiptail Invicta progress is (still) stalled
• No need to rehash Violin

And in purgatory are the vendors who don’t have the same market reach as the winners, but probably will excel at some use cases (some of these are hybrid flash/disk):

• Coho
• Tintri
• Tegile
• Kaminario
• Nexenta
• Nimble

So what’s the differentiation among the current winners?

• The architectures are different for one. Pure is scale up (optimized for latency/scale) where as XtremIO is scale out (optimized for redundancy). Good summary by EMC’s Chad Sakac in this blog post about storage architectures.
• SolidFire uniquely lets you specify IOPS by volume. I’ve got to think they’re working on VVOL integration, which may put Tintri in a tough spot.
• But customers I’ve talked to say it ultimately comes down to which vendor gets the POC first, and the price / performance trade-off. Hence de-dupe, compression, and longevity features matter.

My take? This diversity is ultimately about decoupling price of performance from the price of capacity. You want capacity? Get Isilon. You want performance? Get Violin. You want to tune the combinations? See above vendors.