In 1989, the Loma Prieta Earthquake put thousands of Northern California companies out of business for days. In 1995, dust-like nylon fibers ignited, causing a fire that completely destroyed the $450 million Malden Mills textile mill in Lawrence, Mass. Any type of a disasterbe it brought on by man or by the extremes of naturecan bring business to a halt if you're not prepared for the inevitable. However, you can take three steps now to make sure your data center's resources will continue to function during and after a major disaster.
Step 1: Turn the Data Center Into a Fortress
To make sure the data center won't go down, install multiple power supplies and multiple communication links. Prepare a thorough backup plan. A copy of all your backups should be kept off-site, in addition to the set kept on-site. The off-site location should be far enough away that the backup copies won't be damaged in a disaster that takes down your data center.
Be sure you have local redundancy, using systems such as RAID storage and uninterruptible power supplies (UPSs). The more redundancy a data center has, the less likely it is that the whole data center will go down. Plus, it will be easier to boot back up if a disaster occurs.
Some document-dependent companies may have a hole in their back-up strategy. Although databases and files kept on magnetic storage (such as RAID systems) are backed up often, document images kept in optical storage usually are not. It's not uncommon for a company to perform multiple daily backups of its RAID systems, yet keep only one copy of the actual documents in an on-site jukebox. Optical storage is immune to many kinds of disasters, but not all.
Step 2: Build an Off-site Mirror Data Center
The next link in the chain of guaranteed up-time is to have multiple data centers. In the unlikely event that one data center goes down, you should have a second mirrored data center in place to keep your enterprise going. This means having two (or more) data centers located far enough away from each other so that both aren't caught in the same disaster.
You can run a backup over a high-speed data line to a distant location. Connect two data centers via a T3 or an ATM network and mirror them, or run a backup over the high-speed line.
Step 3 - Plan for Using the Backup Site
Last, but not least, plan for the scenario in which you must re-create your data center and have your employees restart the business at the alternate site. This process may include having equipment vendors supply replacement gear after a disaster or moving employees to an alternate data center.
Redundancy can provide insurance for the data center and the network, but not for the human element. You still need to plan the logistics of getting employees to and from the site in the event of a disaster such as a hurricane. You also must consider what services to provide for employees' families if the employees are away for several days. Emergency preparedness requires working with local and state agencies to ensure everyone's safety throughout the disaster. //
Elizabeth M. Ferrarini is a free-lance author based in Arlington, Massachusetts.