Eat Your Oatmeal -- And Plan For Catastrophe

We all know and loathe the type -- they floss twice a day, eat plenty of fiber, always remember their anniversaries, and max out their 401(k)s so they can retire at 55. When these paragons go into IT, they're inevitably the killjoys reminding the rest of us that it's not enough to have a disaster recovery plan, you have to actually test it every millennium or so.

Don't ignore them just because they're insufferable. After all, do you want the survival of your business dependent on FEMA getting power restored to your main site? We didn't think so.

If you outsource disaster recovery functions for your data centers, you have an agreement with your provider for testing...right? But outsourcing may not be the future. A recent Gartner report that I found interesting says many large companies are choosing to bring DR back in-house. Why? Mostly the crazy high cost of having short recovery time and recovery point objectives. Other reasons cited were the expense of storing burgeoning amounts of data, inflexible and long contracts, and the need to ship tapes an unacceptably long distance.

It makes sense. As virtualization enables us to reduce the number of physical servers needed to run our data centers, and new power and cooling technologies make it feasible to set up a data center in less-specialized facilities, there's even less reason to skimp on DR. In addition, new DR offerings with a virtual twist are springing up. Last week EVault announced its EVault Unified Recovery platform, which covers both physical and virtual environments and is available as SaaS or as licensed software or a managed service. And VMware's Site Recovery Manager eases use of underutilized hardware as recovery targets.

Bottom line, ignoring DR today makes less sense than ever. If you need help getting started, check out our primer on DR. But however you do it, start. And don't forget to floss.