I joined the Marine Corps right after high school graduation because I was tired of school, and restless. I wanted to do something. So, what does the Marine Corps do with me after my basic training? You guessed it, they put me in a classroom for nine months of rigorous technical training to teach me how to fix radios. The training was very in-depth, and I learned how to find problems with every module in a radio. Any module could be the problem. Any module could bring the box down. Everything was a single point of failure (SPOF) with absolutely no redundancy.
Over 25 years later, with the past 17 dedicated to storage, I am still trying to overcome design flaws. Now, at least, the problem is not with the hardware design. All storage arrays are designed to withstand component failures with redundancy built in, no SPOFs. Most data centers are hallmarks of redundancy, dedicated to providing multiples of everything, including backup power. You will be hard pressed to find a single point of failure in a well-executed IT shop. Where’s the problem?
The flaw is on a much grander scale. There’s nothing wrong with the data center, but rather with the assumption that all those redundancies will protect the business if the entire data center where impacted by a major event. Everyone in IT has heard of a catastrophe due to a fire, flood, or massive storm. Similarly, one need only turn on the news to hear of the damage caused by hackers, or disgruntled employees. Even something as commonplace as a backhoe digging a hole can affect a single data center.
What can you do?
To protect your data, and by extension your business, you must plan for crisis! The importance of the data dictates the level of availability needed. By that I mean, if the data is deemed “mission critical” you must employ a copy service, house it offsite – and if needed, provide the means to make it instantly accessible.
Since all data is not equal in value, there are gradients of protection that are applicable, if at all. I define three categories. For data that has no value upon recovery from a crisis, don’t bother to protect it. For data that must be retained, but not necessarily available, back it up to slower, more cost-effective media. Finally, for data that must be accessible, deploy a high availability solution, such as IBM HyperSwap.
There are many solutions dedicated to the protection and availability of data. The goal is to mitigate risk to the business by leveraging the appropriate copy service technology. Chose a solution that meets the objective within the time to recover and cost parameters that the business dictates. As this chart indicates, the closer the solution is to a Recovery Time / Recovery Point objective of zero, the greater the cost.
It’s time for IT leaders to step forward, educate business executives on risk and put a plan together to protect it in times of crisis. IBM has the expertise and the solutions to help.