Dale Vile, originally published on The Register
Has virtualisation really moved the game forward?
OK, so you’ve virtualised a bunch of servers and saved yourself a bit of money on hardware. Life is a little easier because you no longer have to go through the server procurement and provisioning cycle quite so often to meet new requirements.
But has anything fundamentally changed in the way you manage systems and deliver IT services?
All of our research, suggests that for the majority it hasn’t. Server virtualisation initiatives are great for a one-shot payback, but once you have gone through your estate and consolidated what you can, you may not be hugely better off than before.
Virtual servers need patching, monitoring, managing and troubleshooting, just like physical ones. In fact, some report that the administration and support challenge is actually more difficult because it is all too easy for virtual images to proliferate and create virtual server sprawl.
Plus ça change
Furthermore, many have been caught out by virtualisation’s knock-on impact on the rest of the infrastructure. We frequently hear stories of unanticipated network and storage bottlenecks, as well as of funds not being available for adding more capacity because the necessary upgrades were not budgeted for.
Meanwhile, even our latest research, which we will be discussing over the next few articles, is telling us that many IT departments are still not happy with their ability to keep up with demand and meet service-level expectations.
Whether you look at the ongoing level of overhead and hassle for IT or at how well change requests from the business are handled, the picture today is not that much different from how it looked five years ago, before the virtualisation wave started rolling.
The truth is that the real enablers of IT systems efficiency, effectiveness and flexibility are joined up management and automation.
If you examine how the more successful IT departments differ from others, you usually find that they have paid a lot more attention to integration of management tools and processes, and taken steps to cut down on the manual work involved in routine administration.
These aspects of management are closely linked because you can take automation only so far by working in just one particular area.
If you are going to get into automatic or assisted server provisioning, for example, the last thing you want is to spin up a server that immediately chokes because of lack of network bandwidth, or fails because of a storage constraint.
If you are going to get into dynamic quality of service management on your network, your automation policies had better reflect application dependencies across all the relevant servers, storage devices and applications.
In an ideal world, integration of both tools and management processes would come first to prevent different automation initiatives overlapping and running into each other.
Already, however, we have seen problems occurring where tactical automation initiatives win out. A common scenario at the moment, for example, is server virtualisation management stepping over into networking and storage, and then running into conflict with traditional networking and storage tools and processes.
While it is easy to talk in principle about cleaning up the fragmented and disjointed mess of tools and processes currently in place, it is incredibly hard to pull off.
Depending on the size and complexity of your organisation, you might have a number of operations teams in place – one for storage, one for networking, one for x86 servers, one for Unix servers, one for ERP, one for Microsoft apps, one for Oracle apps, and so on – each with its own set of policies, standard operating procedures and tooling.