In Part 1, we discussed the strategy of moving your office or data center, focusing on how to determine your requirements based on the tolerable downtime during your move.
Now that your strategy is established, we’ll talk about the two most common ways of doing the move. Most moves are a variation of these two methodologies. They may not match your plan exactly, but they do represent the two most common starting points for a move like this.
Option 1: Downtime Friendly
“My organization can take 24-72 hours of downtime”
This seems like the easiest way to move your organization. All the equipment involved will be your existing equipment, which you’re presumably familiar with, and should be able to reconfigure as necessary when you move. In fact, this is often the most complicated and stress filled means of migration, since you’re going to re-install your entire infrastructure through the course of a weekend. Even the best of planners often skip a step or fail to account for the inevitable troubleshooting that occurs during hardware installation or re-installation.
If you’re still committed to this strategy, you need to revisit part 1 and identify your dependency map. This is generally going to look something like the following stack chart.
This downtime friendly approach means that you’ll have to have every dependency met, starting from the bottom of the stack and moving up, before the next layer can be put into place. This means you need your UPS installed before you can configure your data center switching, the storage before the compute stack, so on and so forth.
As you create your dependency map, attribute real world setup time to move each component, reconfigure for the new environment, and move on to the next. Allow yourself some time to troubleshoot items which may not work right the first time.
Many find that once this exercise is done, selective pre-build of the new environment (i.e. new equipment) will significantly reduce the amount of work to be done in a short amount of time.
Option 2: IT Friendly
So you’ve realized that moving everything in a weekend is an admirable goal, but one fraught with potential pitfalls and downtime that is unacceptable to the organization. You’ve decided to pre-stage some or all of the infrastructure using new equipment, giving you the opportunity to design and test the new configurations without the stress of a stopwatch.
Your dependency map still applies. But with this approach, we take the estimations of time, and identify the most lengthy and error prone portions – we’ll target those for replacement first.
Most organizations minimally look at UPS/Power and network components as a baseline of infrastructure to stand up at a new location. These are major dependencies that every application needs. These are also some of the most time consuming elements to configure if you’re moving and reconfiguring switching environments.
There are more than a few strategies around moving services such as websites, applications, active directory and other components. I’ll cover those in the next part of this series.