Hurricane Season Coming: Prepare Your Datacenter

It pays to be prepared.

It’s August, and that means hurricane season for the U.S. East Coast is rapidly approaching. In 2011 and 2012, storms smashed into the region and caused significant damage; a lot of people are hoping that 2013 won’t feature a repeat—especially those with enormous datacenters to keep intact.

When Sandy plowed past New York City in the fall of 2012, it caused widespread flooding and power outages. In a bid to keep their Manhattan datacenters running, various IT managers took extreme measures: Peer1, for example, even set up a bucket brigade that carried fuel up multiple flights of stairs to the generators.

Preparing for the next storm (or other natural disaster, for those not on the East Coast) is no easy task, but a couple of tips can go a long way toward ensuring that data stays safe in the event of calamity:

Do Preventative Work Now: That means not only deciding on whether to initiate a backup datacenter (hopefully in a different geographical region), but ensuring there’ll be personnel to actually reach and maintain that datacenter if disaster strikes. Any number of companies choose to enlist a big IT vendor to handle backup duties, which spares them from having to build out their own infrastructure (and employees)—just remember that certain needs (zero data loss, instantaneous rollover, etc.) dictate the infrastructure, and not all third-party vendors can provide all types of support.

Decide What’s Critical… and What’s Not: If your datacenter goes down, what can you afford to lose in terms of data? What datasets and databases can lie dormant for days, weeks, or even months without any interruption to the larger business? Along those same lines, what can you port safely to the cloud, and what do you need to leave hosted in an on-premises datacenter? The answers to those questions will dictate a company’s backup plans.

Security Is a Huge Issue: For security-sensitive industries such as finance, a backup location needs the same protective measures in place as a primary datacenter. “During disaster is when companies are the most vulnerable,” Richard Cocchiara, an IBM distinguished engineer for business resilience, said in an interview. “When people are distracted, and the company infrastructure is under stress, is when criminals try to take advantage.”

Remember the Network: In the end, it’s not just about a single datacenter, or even the backups. Any company that wants to sail through a disaster with a minimum of collateral damage will need to ensure that data is secure and safe both in movement and at rest.

Authority: When hurricane Sandy brushed New York City, any number of IT administrators had problems figuring out who to call when they needed authorization for certain actions, such as shutting down a datacenter or moving data from one location to another. Limited phone and email service made communication even more difficult. In light of situations like that, Cocchiara recommends that companies establish a chain of command beforehand: “There are things you want to automate in the absence of people being able to communicate.” If a CIO is unavailable, someone in the middle management tier may need to make a huge decision—choose that someone long before disaster strikes.

Good luck.


Image: hxdbzxy/