A couple of months ago I attended VCE’s first annual analyst conference. For those that don’t know VCE, it’s a joint venture — primarily between Cisco and EMC, with a bit of funding tossed in from VMWare to round out the offering. VCE acts as both a technical integrator and a professional services firm for the converged cloud solution.
I’ve talked to a number of VCE customers and almost all of them have told me that converged solutions helped minimize the deployment complexity of implementing a private cloud solution. This is why so many vendors have focused on building their own converged infrastructure solution. One could make the argument that the decision to use converged infrastructure is a no-brainer for any organization looking to turn up a private cloud. Seems obvious: need a cloud, drop a converged solution in and off you go — easy, peasy, lemon squeezy.
On paper the decision may seem obvious, but in practicality the shift to converged infrastructure may not be as simple as it seems. Not because of any technical issues, but more because of operational structure. I believe any company looking to deploy a private cloud needs to shake up how the IT organization is structured.
Typically IT departments are organized along technology lines. There’s a storage group, server group, virtualization group, network group, etc. and they all work independently to some degree. The lack of communications within these groups is one reason why IT projects often take so long. Years ago I was consulting at a very large organization — a brand everyone would know and have first-hand experience with. All I wanted was a rack-mount server running Apache to try out a new web application. Well, the server was being ordered from Dell so that took six weeks alone. The person who allocated rack space wouldn’t assign me a spot in the rack until the server came in. The person running cables wouldn’t even consider doing it until the server was in the rack. All in all, it took six months to get the server up and running. Hardly the on-demand environment businesses seek today.
A private cloud would certainly streamline this. One of the value propositions of a private cloud is having the ability to self-provision infrastructure. Had my organization been running a VCE (or other converged) solution, I could have easily created a virtual server and loaded the web server on it myself.
In this type of environment, the traditional structure for IT makes no sense at all. Storage, server, and networks aren’t managed independently. Instead IT needs to fall along different lines. New operational functions would include the infrastructure itself, variation of hypervisor (Hyper-V versus ESX, etc.), application type, etc. While this may sound good on paper, making such a change isn’t easy.
I recently interviewed the CIO at a large university about their deployment of converged infrastructure and private cloud implementation. He told me that, by far, the biggest challenge he had in shifting to a cloud was organizational issues. Most of his staff had been in their role for some time and, as the old axiom goes, if it ain’t broke, don’t fix it. At first, he thought the groups could work things out themselves but he quickly realized that was a pipe dream and was forced to mandate change, which included moving some people out and bringing new people in.
It’s with this in mind I issue the following warning: IT is changing whether we want it to or not. Those that resist and want to keep things the old way will soon find themselves on the outside looking in. CIOs that don’t mandate the change will find keeping up with competition becoming more and more difficult. Infrastructure is changing and the way we manage IT must change along with it.