A few weeks back I wrote about what Disaster Recovery (DR) means to an organization and put forward the suggestion that there is not always enough importance placed on adopting and deploying a proper DR strategy.
This post will serve as a follow up, and I’ll outline some best practices for planning a proper DR strategy that’s more like a substantial and nutritious main order.
So what is the recipe for a great DR plan?
Here’s my 7-step recipe:
1. Define your recovery point objectives (RPOs) and recovery time objectives (RTOs). How much data can you afford to lose? Be realistic! Aggressive RPOs and RTO’s tend to have a more expensive price tag when considering the replication method, and may be overkill.
2. Identify the critical services or applications in your environment. These will ultimately tie back into your RPOs\RTOs. Evaluate whether they are truly critical or not. It’s not always realistic to put everything in the “business-critical” basket; this can lead to bigger data sets to replicate, which can mean a more expensive replication technology.
3. Plan your recovery. Create a run book; if you are a CIO, then work together with your lead architect to understand the technical AND business implications of what the run book needs to look like. Applications have different change rates, so there needs to be a collaborative approach from the technical and business owners.
4. Test and time the run book. A plan is no good unless it is often tested and updated. All too often, DR strategies are implemented but not thoroughly tested. Again, you don’t want to find out your DR strategy doesn’t work at the time of an actual disaster. You want to know beforehand – otherwise, you might end up running around like a headless chicken. Consider that things are always changing in IT environments, and these changes could mean your recovery plan is no longer valid.
5. Optimise and plan again. Are there any steps that can be enhanced or removed? Perhaps you can research an automated software platform for managing the failover process.
6. Leverage virtualization where possible. The benefits of virtualization adoption are already a well-known concept in today’s industry; it places less reliance on moving and connecting servers in disaster scenarios.
There are some very impressive DR automation software packages, which are designed to automate the DR process to give the business the confidence that the recovery process will work, eliminating operator errors and reducing reliance on experts. These tools can aid and support a solid DR solution by orchestrating the virtualization recovery plan according to the defined business metrics (RPO\RTO). A fully automated DR strategy is not necessarily the best path; more automation can mean faster recovery of service, but may limit the testing capability.
7. Invest in a solid and proven replication technology. This can be either host-based or array-based, but again, it always needs to be linked back to the defined RPOs and RTOs of each of the services. Elements such as the overall size of the environment and numbers and types of applications to support will begin to provide some direction to the kinds of replication solutions that are likely to fit.
Appropriate replication options should be evaluated, selected and driven by the business and the SLAs that were initially defined.
Learn how HP Storage Technology Services can help you improve business continuity.