It never hurts to be prepared for the worst-case scenario. For smaller hospitals that rely on a single data center, backing up data so they can be recovered may not be enough if the data center is down for any length of time. Leaders at Beaufort Memorial Hospital in Beaufort, S.C., took that into account when developing a disaster recovery plan, says Ed Ricks, CIO.
A 197-bed community hospital that serves over 100,000 people, Beaufort Memorial is located near the Atlantic coast, making it vulnerable to hurricanes and floods. And while Ricks says his disaster recovery plan has “morphed over the years,” he emphasizes that in addition to its in-house data center, the hospital also rents space in a building further inland that serves as a “warm” site.
In the past, as real-time data was replicated, the data were sent to the warm site where it was stored. But because that warm site was not an active site, it lacked the ability to run any of the hospital’s information systems.
However, since the hospital had already approved another medical arts building on the campus, Ricks says the decision was made to put a second data center in the new building. “That’s now our secondary data center but it will soon become our primary center,” he explains. “We want it to be an active-active data center.”
Ricks says the hospital partnered with Mt. Pleasant, S.C.-based eGroup to design a plan that would incorporate two data centers instead of one. In addition, Beaufort Memorial had already set up its backup storage solutions with the Hopkinton, Mass.-based EMC Corp., so EMC was brought back to work on the expansion project.
Now, says Ricks, the two data centers are linked with a fiber optic cable owned by the hospital with a current capacity of 10 gigabytes per second. The warm site also is connected, but by a leased line that is part of the backbone for the state’s health information exchange (HIE), he says.
Ultimately, Ricks says, he would like to turn the warm site hot in case both on-site data centers ever go down. “The long view is to turn the warm site into another active data center — probably in a few years.”
In backing up data to disk, one copy is now stored in the secondary data center while a second copy is stored at the warm site, he says.
Working with eGroup was fortuitous for Beaufort Memorial, since that applications and services provider has strategic partnerships with EMC, Cisco Systems, and the Palo Alto, Calif.-based VMware.
Beaufort Memorial continues to add new technologies to its arsenal: not only does it no longer use paper charts, but it also added a virtualization platform using VMware’s vSphere, Ricks says. “We run 225 servers and about 95 percent are virtual.”
Developing a disaster recovery strategy and building in redundancy with a second data center was not without its challenges. “You can do anything you want to do with enough time and money. But you’re always looking at competing dollars,” he says, noting that IT expenditures are often at odds with those on the clinical side.
As for offering advice to other CIOs who are planning to implement a disaster recovery strategy, Ricks urges, “Have a long-term plan but make sure you have an end point in mind. Communicate what you want to do, get everything in place, then test it.”
Plus, he notes, “It’s not a luxury; it’s a necessity.”
