No one ever got fired for buying IBM. Well…how did that work out?
When I started working in storage, it was a major portion of our capital budget. When we made a decision on a storage platform, we had to write the proposal for the CIO to change to another brand, and we had better be sure we didn’t have issues on the new platform. We didn’t buy on price, we bought on brand, period.
I was speaking with a customer recently, and they were talking about how they were moving to a storage startup which recently went through an IPO. I asked them how happy they were about it, and the response was, something to the effect, it is great, but we will likely make a change in a few years when someone comes out with something new and cool. This wasn’t an smb account, not a startup, this was a major healthcare account. They were moving away from a major enterprise storage vendor, and they were not the first one I had spoken to who is going down this path.
I remember when virtualization really started to take off. The concept was amazing, we thought we were going to see massive reduction in data-centers and physical servers. Please raise your hand if you have less physical servers than you did 10 years ago. Maybe you do, but for the most part I rarely see that anyone has significantly reduced the number of workloads. I guess virtualization failed and was a bad idea, time to move on to something else? Of course not, we just got more efficient and started to run more workloads on the same number of systems. We got more efficient and better at what we do, we prevented server sprawl, and thus realized cost savings through cost avoidance. What has changed though is moving from one server vendor to another is pretty simple.
If I were still in the business of running datacenters I would probably spread over two or more vendors with some standard builds to keep costs down, and provide better availability. From a storage perspective I wouldn’t really care who my storage vendors were provided they could meet my requirements. Honestly I would probably build a patchwork datacenter. Sure it would be a bit more work with patching and such, but if there are API’s, and we can do centralized management to deploy firmware to each system, why not, why be loyal. For that matter, why have a single switch vendor?
See what I did there? It is all about the software. Whether you believe VMware, Microsoft, Red Hat, or someone else will win, the reality is it is a software world. If your hardware will play nice with my hypervisor, and my management tool, why should I use only one vendor, if it won’t, why should I use it? It is all about applications and portability. Hardware isn’t going away, but it is sure getting dumber, as it should, and we are pushing more value through software. He who controls the management software controls the universe.