So one of my colleagues is working at an investment bank in the city. They’re about to embark on a virtualization project. The first step is to virtualize the hardware, to get rid of the boxes, once that’s done the business lines can be asked with the infrastructure teams to think about consolidating their applications.
The question they’re having is how to do this? Do we outsource it to vmware/hp/ibm? Do we insource it, get some guys in, and hand it to them plus our internal teams? There are benefits and costs to all these options:
Outsourcing problems:
- Lack knowledge of the infrastructure and applications
- Will cost more than getting someone to do it internally – they’re already paid for
- Production support might feel they aren’t involved/able to support the platform
Outsourcing benefits:
- It’s someone elses’ problem
- They’ve done it before and know the issues
- You have support/escalation if there’s a problem
- They can handle the initial support problems
- Might be cheaper – one off fixed cost Insourcing problems:
- Might be dependent on contractors that could leave during project
- Might take longer to get started Will be made to fit within the current political/”the way we do it” framework Expensive – dedicate a few engineers time to this?
Insourcing benefits:
- You get the control Application teams can speak directly to engineers they deal with daily
- Can be made to fit within the way you do business on your terms
- You can do things to fix problems that might not be officially supported
- Use your existing relationships to get things done – avoid that’s the way “vendor” does it T
There’s a tendency is for the banks/companies to own the process themselves, get the guys in they need and proceed.
Yes there are the usual issues in implementing new technologies, but I can better manage the process. I know what the expectations are, I know “what I can get away with”, who to speak to when problems occur, how to manage failure within our organization, and if planned correctly, it can be used as a relationship building exercise, “We want to virtualize your infrastructure, let’s talk about you and your application”; often the infrastructure teams don’t know much about the applications, this gives them a chance to learn, it makes the infrastructure meet and work with the application teams.
You’ll have the usual challenges, but if handled correctly the technology can be used to show how effective IT can be, that on-demand computing can be achieved, and can allow you to force all those cross charging issues to be resolved.
No related posts.
Related posts brought to you by Yet Another Related Posts Plugin.

Blog videos
Quite correct in most respects, all boils down to correctly profiling the existing applications in the first place to make sure that they are suitable for virtualisation. Time and time again, developers deny that their existing servers require much power, yet when virtualisaiton is mentioned, the request for a single virtualisation guest size expands until it is better to provide a seperate physical machine.
[...] on the train to London), so I haven’t covered everything. Related to this are the articles (one and two) that were written in the past – note the process and operations are just as [...]