A week of planned–and unplanned–natural disasters is keeping data center professionals on their toes.
By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
First, a 5.9 magnitude earthquake hits the East Coast on Tuesday, causing little damage, but rattling a lot of nerves. Next up, Hurricane Irene is predicted to make landfall in the mid-Atlantic states over the weekend as a Category 2 storm, threatening data center-rich areas, such as the Carolinas, Virginia, Washington D.C., New York, New Jersey and Boston.
As they wait for Irene to hit, East Coast data center pros are revisiting disaster recovery plans and keeping their fingers crossed.
“We’re doing a lot of blocking and tackling”–things like checking available bandwidth and VPN capacity, and revisiting support models in the event that people can’t make it in to the office, said James Duffy, vice president and enterprise architect for Advanced Health Media, a sales force logistics and compliance management provider located in New Providence, N.J.
On the facilities side, data center operators are doing the equivalent of buying bottled water and bringing in the lawn furniture.
“We touched base with our diesel vendors and topped up our generators. We walked the roofs to make sure there was no loose flashing and checked the latches on our air handling units,” said William Moore, CTO at CareCore National, a benefits management firm with vulnerable data centers in Atlanta, South Carolina and New Jersey.
But when it comes to maintaining system uptime in the event of a hurricane, a lot comes down to upfront system design.
“We’re not doing anything differently than we normally do,” said Bill Akins, a systems specialist at Emory Healthcare, based in Atlanta, Ga. With four sites and applications spread across and replicated using EMC Corp.’s SRDF, the hospital system is already designed for uptime.
Similarly, CareCore is putting its faith in its highly-virtualized, geographically-distributed infrastructure architecture. The firm runs a 100% virtualized environment on a series of VCE Company’s Vblocks– converged infrastructure pods that combine Cisco Systems’ servers and networking, and EMC’s storage –in its four data centers, and load balances between them.
Up front, “we validated that any one data center can assume 100% of the load,” said Matt Cunningham, CareCore senior vice president for IT. The firm doesn’t plan on moving workloads preemptively, since much of its load balancing is done automatically. But if push comes to shove and a facility becomes unavailable, “we can move a data center simply by moving workloads from one set of infrastructure to another,” he said.
Other hurricane-weary firms aren’t so sanguine. Woodforest National Bank, for instance, has a primary data center in Woodlands, Texas, but moves all its production applications to a distant colocation facility for four to six months of the year, before the start of hurricane season, said Richard Ferrara, CTO. It’s an expensive way to go, he acknowledged, but, “bottom line, you’re not running around trying to get things moved if a hurricane is coming.”