HPE is acquiring Nimble Storage for $1.09 billion. Now that caught me by surprise! Before my time at Tegile, I was an HP employee (they weren’t HPE yet back then) by means of their acquisition of 3PAR, so I have some unique perspectives that I would like to share and would love your comments on below.
As a competitor of Nimble, I have to admit that in the last few years they have done a nice job capitalizing on the shift from legacy storage to flash-based storage systems and built a healthy storage business. Nimble’s portfolio of hybrid and all-flash arrays are good solutions in the mid-market. So, my gut reaction when I heard about the acquisition was that HP now has a good substitute for their aging Store Virtual line. Nimble’s hybrid arrays will deliver good value in the mid-market where Store Virtual (aka Lefthand) played. But as I thought more about Nimble’s all-flash offerings, I was left scratching my head as to why HPE would buy Nimble given that HPE had re-engineered 3PAR products to effectively play in the sub-$20,000 price band.
This is the part that confuses me unless my ASIC conspiracy theory is true. You see, the heart of 3PAR’s StorServ architecture is based on an ASIC. It makes the systems run screaming fast – even with disk drives, and lets the system very gracefully scale up to 8 nodes. Cool stuff. The rough part about ASIC is that it takes a very long and expensive development cycle to turn new features into an ASIC. I know that many HP engineers were trying to design that ASIC out of the system to get development costs down and agility up. Perhaps they couldn’t do it, so on the lower end, where features get added very quickly, HPE needed a different platform to add features to faster than an ASIC turn, while leaving 3PAR for the big iron customers that they only sold to before being acquired.
The decision to keep or dump the all-flash systems from Nimble is less clear to me. The 3PAR all-flash product, with all its bells and whistles, is a significantly better platform for high performance, low-latency, business-critical workloads than Nimble’s all-flash line – scale out and federation, quality of service and 3-way synchronous replication features clearly set 3PAR all-flash offerings apart from Nimble’s all-flash portfolio. But Nimble’s all-flash in conjunction with InfoSight is far easier to use and manage compared to that of 3PAR. So it’s a bit of a toss up.
That brings me to Nimble InfoSight. Did HPE buy Nimble for InfoSight? Nimble made a big bet hitching their value proposition on InfoSight. Almost all the feedback I hear is that InfoSight is pretty cool. (I’d be remiss if I didn’t tell you that Tegile’s IntelliCare Cloud Analytics is very cool as well.) 3PAR has a thing called STaTS. It collects tons of data about the arrays, but the interface is (at least it was when I left) extremely challenging to use. I can certainly see HPE porting STaTS data into InfoSight for a federated cloud analytics layer across the two lines. That being said, $1 billion is a lot of coin for an analytics platform. I am sure the sales team at Glassbeam would be happy to build you a machine log analytics platform for a billion dollars, right? So, it must have been for more than just InfoSight.
The other question I ask myself is: “Can HPE’s channel get nimble?” I doubt it. That part of HPE is a big cumbersome beast with enough back-end rebates wrapped around themselves, it would make your head spin. A big part of Nimble’s success was their execution in the channel (great job guys – I really mean it). We’ll see how this shakes out, but there are several customers and resellers on reddit that seem awfully worried.
There are lots of things changing in storage these days, this being the latest. Whatever happens, I am sure it will be a fun ride for all of my friends that work(ed) at Nimble and HPE.