Nutanix’s IPO launches hyper-converged but can it escape the rules of the storage universe?

We’ve been tracking the rise of hyper-converged infrastructure recently, and one key measure that seems to confirm this as a rising star of the tech world is Nutanix’s IPO last week.

Its shares were initially offered at $16 but these rocketed as soon as they hit the market and went as high as $44.46 this week before settling back a little.

Such performance is a vote of confidence in Nutanix and in hyper-converged infrastructure.

So, what’s the thing with hyper-converged?

Well, if you think about it it’s a relatively simple idea. You take server hardware and put it in the same box as some storage.

So far, so . . . well, like a “traditional” server really.

But there’s more, and here’s the key(s). #1 The server capability is virtualised, out of the box, so customers can get straight on to delivering apps via virtual machines. #2 The whole package comes in nodes that can be linked in scale-out fashion, so as capacity and performance needs increase you can add more.

Add all that together and you get a very customer-friendly package, that’s easy to deploy and is pre-configured so its different elements work together without any inter-vendor head-knocking to keep IT chiefs awake at night.

Limitations? Well, currently hyper-converged seems to be pretty much an SME or branch/remote office play. Although clearly capable of supporting fairly demanding virtualised environments, it isn’t generally used as high end transactional storage.

Still, we should expect hyper-converged to make mincemeat of entry-level and some midrange storage going forward.

But where next for hyper-converged? More flash, more software-only options, greater ability for regular (ie, not webscale giants) to build something more like hyper-scale environments and to nudge towards high end transactional environments.

Therein, however, possibly lie the seeds of the current vendor-led variants’ challenges.

Hyperscale, which is arguably the inspiration for hyper-converged, was/is effectively a DIY phenomenon. At root all storage/compute has commodity kit at its heart and the tension is always there between the commodity vehicle and the software component that defines the product.

In that, hyper-converged cannot escape the laws of the storage universe that are slowly eroding the rule of the incumbent giants of the market.