Three steps to a successful data center migration plan
Getting pushback from customers to do it faster and cheaper without increasing downtime? Walk them through the data center applications and ask them how long they think the migration should take and what the consequences of downtime really are. Can they afford the risk of faster and cheaper?
Step 2: Prove that your detailed plan supports service-level agreements (SLAs).
A successful data center migration plan is based upon a solid migration plan that factors in overall program management, IT infrastructure, facilities planning and coordination, application migration and lots of contingency planning.
The hard part is proving that the application won’t break during the migration. Breaking an application means experiencing degraded performance or downtime outside of acceptable SLA limits. Creating a contingency plan for the application migration is essential to winning the confidence of the business so that the migration can take place. The more comfortable application owners are with the plan (and recovery), the more cooperative they will be. Their participation is critical from both a technical and functional perspective. You will need their help in the following areas:
- Understanding the application architecture and dependencies
- Validating the applications maps and inventories
- Testing the application once it’s been moved
- Understanding planned application releases and upgrades
- Listing restrictions for peak seasonal activity, if any
- Identifying customer SLAs that must be met
- Quantifying the business impact of downtime
Proving that your plan mitigates application downtime risk to the business is critical to the success of your data center migration plan.
Step 3: Test and report on success and failure.
Testing and reporting are critical components of a data center migration plan but are often understaffed and poorly funded — at least until something breaks. Solid testing and reporting offer the best insurance against production breakage. You just have to do it.
The best planning in the world does not remedy poor execution. Excellent execution does not remove the need for testing. And testing twice is better than testing once.
Organizations that regularly migrate applications rely on solid staging, testing and reporting mechanisms to ensure they minimize application disruption (downtime). They know the ultimate success measures of a data center migration are staying within budget, staying on time and minimizing disruption. It’s important to report testing results with application owners and highlight any problems you find so that you can jointly resolve them prior to the actual migration.
This was last published in September 2008