In a perfect world, recovering a company's data and getting it back into production after a major disaster would be fast and automatic. Got some news for you: Even "good enough" is rare in this world, never mind "perfect."
First of all, industry analysts from Gartner and IDC say that 30 to 40 percent of all IT shops either have no disaster recovery system in place or do not know how to use it correctly. Second, even if a shop does have a DR apparatus in place and tests it occasionally, there are plenty of examples of such systems not performing according to plan.
In the world of data recovery, the data is the easy part, the recovery can be hellish, and IT administrators are the ones commissioned with connecting the dots. Enterprises laid up for extensive periods of time due to IT knockouts do not have a glittering record of surviving, so there's more than a modicum of pressure involved here. The National Archives and Records Administration reported in 2010 that 93 percent of enterprises whose data centers were down for 10 days or more due to a disaster filed for bankruptcy within one year of the disaster.
The term "data recovery disaster" defines a situation in which an extended power outage forces an enterprise to recover its data and files from an alternate location-whether that is within the enterprise's physical system or in a cloud backup and recovery service. A short-term power outage or failure of an individual server, or even a rack of servers, isn't generally considered a data recovery disaster.
Hurricane Katrina in August 2005, the March 2011 earthquake off the coast of Japan, and the recent flooding in the Midwest are well-known examples of multidimensional disasters that affected scores of IT systems. The most common cause of IT failure? By far, it's human error, whether involving a power supply or IT production itself.
What are some important general DR guidelines for an enterprise-any size enterprise? They're simple, at least in concept:
- Find a system that fits your business and implement it. Don't laugh; many companies don't have one.
- Select a system that includes snapshots, mirroring and/or replication to a separate location, whether that location is within the confines of the physical enterprise or a cloud-service package.
- Test the system on a regular basis, even if it involves just a portion of the system at a time.
DR systems vary greatly. Along with major advancements in the last five to six years in system storage capacity and overall data protection, there have been great advancements in recovery methodology.
It should be noted that backup storage alone is not a DR plan. Most companies do back up their data, but that doesn't cover getting applications, networking, cloud services and everything else back up and running with the stored data.
If your company's DR plan consists of backing up files every so often and then yelling "Help!" to your IT department or a service provider when an outage occurs, then you need a real DR plan.