At the application server level, where we run VMware and Windows on Dell servers, the content of each virtual machine is also replicated to the disaster recovery site anytime an update occurs. With VMware and IBM database servers, we use a set of servers for testing and development. When we need to run a disaster recovery restore, we turn off the virtual servers for test and development, bring-up the ones for disaster recovery, and we're good to go. All the data and content of the servers is quickly copied over.

Four-Step Disaster Recovery Process

For enterprises ready to develop a disaster recovery plan, we recommend a four-step process that helps frame the project and ensures a reliable disaster recovery process:

Step 1: Enablement
Make sure all the data is properly transferring to the disaster recovery data center. Ensure that all the proper hardware in the disaster recovery data center is in place, will remain stable and is running on up-to-date operating systems. Also, review all applications and decide how long you can you go without each one. This helps prioritize the most crucial applications. Some applications might need to be restored in less than an hour while you might be able to do without others for up to 12 hours. This part of the plan becomes an internal SLA.

Step 2: Testing
Develop detailed procedures and processes on how and how often to test the disaster recovery plan. We recommend at least once per quarter. You also need to determine how to measure success so that you can evaluate the testing and document the findings to compare one test to another with a high level of validity.

Step 3: Cutover Documentation
You need to document exactly how you will cut over if and when a disaster strikes. There will be some elements similar to the test process, but there will also be differences for how you execute procedures while under a live disaster recovery. With all the pressure your IT staff will be under, it's critical that this step be clearly and thoroughly documented.

Step 4: Returning to Normal Production Infrastructure
Just as important as how to cut over to your disaster recovery infrastructure is knowing how to return to your normal production infrastructure. It's not always a case of doing things in reverse, and it's a process you should also test.

Lessons Learned

It's important to bring all of the key vendors and your internal IT team into the same room at the same time. This gives everyone a chance to voice concerns, explain how their piece of the puzzle contributes to the overall project, and to understand the functions of the other parts of the project. If you get yourself into a position where you act as the go-between among your vendors, important information will undoubtedly be lost in translation.