Just like the real world, there are various types of marriages in the storage ecosystem. Some are arranged (“Please buy us before we run out of VC money!”), some are expected (“As a systems behemoth, we need to acquire your cool new technology…”), and some are enjoyable emotional battles (“There’s no way I’m gonna let you hook up with him/her, so here’s a larger dowry…”). Every so often there’s one that doesn’t fit the standard mould, but actually when you look at it, it’s easy enough to see how the partners fit together.
And so, dearly beloved, we are here to bear testimony to yesterday's joining of Fusion-io and NexGen (well, of course the former actually bought the latter for $119m cash/stock, but let’s not spoil the story on such a brash technicality!). So what do we see for the happy couple?
Fusion-io brings – of course – great capabilities and success in the server-flash world. Having started with ‘straightforward’ on board server flash (ie DAS), the vendor has made strides over the last couple of years to overcome the obvious number one drawback of that approach (the performance advantages are equally obvious!) by adding a couple of ways to get the relatively (as compared to spinning disk) expensive flash shared across systems. You could do this via its all-flash approach (shared across servers and apps as needed) or via making the in-server flash shareable across storage systems (this is the caching option that it got by buying ioTurbine). These various options are all-well-and-good but they are pretty much focused on solving known issues in given workload areas.
Meantime we have NexGen, which stated ‘dating’ Fusion-io in 2010, and built its hybrid storage system around the ioDrive, and the abilities it provided. Netting out the NexGen model, it is an SMB-focused system that is aimed at serving a broad – and dynamic – range of workloads for users that don’t have armies of white-coated sys admins available to continually tweak things. Other players have similar sounding attributes but NexGen’s claim to fame is an ability to deliver performance to individual apps according to what their priority is to the user….not QoS in the responsive sense, but QoS in a deterministic sense.
Sounds good eh? Here’s the challenge – there are already hybrid storage systems with traction (whether traditional players such as the EMCs of the world or emerging ones like Nimble). Much as there are indeed significant differences between all the options, it’s really hard to get these noticed without investing heavily in marketing, sales, and support coverage, and further R&D investment. Meantime Fusion-io knows a couple of things: first, that its total available market (TAM) – while good - is limited by its high-end focus; second, that – unless and until all storage costs the same – there will always be a need for a storage hierarchy.
And so there, as they say, you have it.
NexGen can get to market faster and broader by embracing its specialist suitor, and Fusion-io just got access to a much broader TAM. NexGen can ramp without so many issues around financial viability or start-up FUD; and Fusion-io has opened up its piggy-bank to move further away from its component roots by buying into the general storage systems market.
It looks like an interesting marriage for sure…but something tells me the ‘neighbors’ aren’t going to be quite so cheerful as and when the couple tries to move into their backyard.