Executing a Physical To Virtual Transition

Over the last couple of days I’ve been discussing some of the aspects of a project I’m involved with. Today is the third and final part. I’m now planning the transition of physical systems to the new virtual environment that we’ll be installing.

The systems we’re planning to move run on a variety of different hardware including desktop computers and server hardware. As well as some industry-specific applications, we’ve got a few of the usual suspects such as email, payroll, finance and a project management tool we are using for a major infrastructure project the company is engaged in.

The applications run on a variety of different operating systems including Windows 7, Windows Small Business Server and Linux.

One of the main goals of the transition is to move things with no interruption to normal operations. That doesn’t mean no downtime – but it does mean migrations of some applications will need to take into account specific business cycles. For example, the payroll system can be offline for a day or two – as long as it does not coincide with the fortnightly pay run.

Once we plan the order and timing for each application’s migration we then intend to set up all of the virtual servers we need and install the required operating systems. Once they’re all running, hardened and tested, we’ll install the required applications on each server and migrate data for the first time. This will give us an opportunity to test applications and get sign off from each application owner.

Once we validate our process, we’ll schedule the migration of production systems in accordance with the least risk to business operations.

What do you think of this approach? What would you do differently?