Brian Jackson - Fotolia
Often, I hear people discuss the concept of vendor lock-in as if it's avoidable. Sure, to some extent, that may be possible, but there comes a point at which you'll need to make a decision and run with it. And, frankly, every decision you make results in lock-in to some degree.
It's a myth that lock-in is completely avoidable, including the risk of hyper-converged vendor lock-in. The fact is you're going to be locked into a platform, whatever it is, for at least your depreciation period, unless something really unexpected happens.
Of course, with hyper-converged infrastructure (HCI), you're locking in three resources all at once -- servers, hypervisor and storage. So, it would seem, what's beneficial about HCI effectively triples the risk of hyper-converged vendor lock-in over, say, buying servers, a hypervisor and storage separately. This often leads IT decision-makers -- when trying to decide how to move forward with an infrastructure purchase or upgrade -- to lament, "But if I go with a hyper-converged infrastructure vendor, I'm effectively buying lock-in for the whole IT infrastructure."
Or are they?
The real hyper-converged vendor lock-in risk
Today, a lot of hyper-converged platforms are software-based, so you get to choose whatever server you like. You can even choose to go with a prebuilt appliance or with options from a vendor's hardware compatibility list.
For the hypervisor, a number of hyper-converged offerings support multiple hypervisors, and, at a minimum, most support vSphere, which still leads the hypervisor market. If you choose an option that doesn't use vSphere, there are often other business or technical drivers pushing you in that direction, so I wouldn't consider that lock-in, necessarily.
Next up is storage. Yes, you have to use the storage services offered by the hyper-converged platform. So, again, this might be considered lock-in. And, as you add nodes, you need to buy licenses or hardware from the HCI vendor. Lock-in everywhere?
The fact is it's really not. Bear in mind, you're making a conscious decision around all of this, and, in many cases, there are ample opportunities for choice. Moreover, let's say you purchase an HCI product and discover in six months it's really not working out. On the technical front, hyper-converged platforms run virtual machines using software-defined storage. It's not difficult to migrate workloads off these systems to other platforms should it become necessary.
Hyper-converged financial lock-in
Perhaps the key issue isn't hardware or software lock-in, but financial lock-in. If you have a three-year replacement cycle and you buy an HCI product that gives you six months before you realize it's the wrong tool, then lock-in comes down to making a financial decision, not a technical one.
The question becomes: Does it make sense financially for you to get out early, or do you wait it out? This is how I've often considered the whole lock-in concern over the years.
At some point, you have to make a decision. I realize there can be other factors, but you're not being forced down a particular path. If you go in willingly only to find you've made a grave error, the real risk of hyper-converged vendor lock-in revolves more around money, not technical limitations.