Electric Lobby Turns On Failover

When your business is lobbying, constant access to email is critical.

That realization prompted Washington-based Edison Electric Institute, an advocacy group for electric companies, to put an automated, high-availability system in place for Exchange. Lead network engineer Rick Williams says that decision saved him a major headache soon after implementation.

"Email is extremely critical to the type of late night, after-hours work we do," says Rick Williams, lead network engineer. "We have a hard time performing our core business without email."

That core business consists of providing analysis and other information for shareholders of electric utilities around the world. EEI analysts constantly share information with other members of the organization, Congress, government agencies, and financial firms.

Until last year, EEI was mirroring Exchange servers between Xiotech Magnitude 3D SANs at the organization's Washington headquarters and a disaster recovery site about 25 miles away in Dulles, Va. "That allows you to recover and get back up if there's a problem, but it's by no means a quick thing," he says. "It takes a while to reposition servers. It's time-consuming, and takes from around four to eight hours to recover."Williams wanted an application that could run on the SANs and automatically failover and failback his Exchange server in case of a problem. After looking at applications from vendors he'd rather not name, he settled on XOSoft's WANSyncHA (high availability) for Exchange. (See XOSoft Readies Replication.)

He says WANSync cost more than some of the other applications he considered, but he liked the bandwidth throttling, which buffers data until WAN bandwidth is available, as well as the compression capabilities.

"We wanted a small footprint that wouldn't affect other services running over our Exchange server. We didn't want the cure to be worse than the disease. It wasn't the cheapest product [at around $10,000], but it fit for us."

EEIinstalled WANSync last August. A month later it came in handy when Williams learned that Verizon would knock out the power to EEI's building for an entire weekend for maintenance on the phone system. At the time, the EEI had analysts at an industry conference in Arizona. "So that was exciting news," Williams says, and he doesn't mean the good kind of exciting.

"People would be mailing a lot of critical information back and forth. We knew we were going to have to be ready to failover because these guys were going to be working through the weekend and we had to have email in place."Williams says he was able to failover Exchange to his disaster-recovery site on Friday night, hours before losing power. "In 20 seconds we were back up. Our people in Arizona didn't even know we had a power failure. It was what we expected, but sometimes you never know until you do it."

Failing back over when power was restored took a bit longer, but went without problems. "When we switch back to the master server, the replica is still your email server until it synchs over and catches back up. It took a few hours, but during the whole time, the application sends you plenty of system messages and warnings and tells you what the process is."

He hasn't installed the other modules, but plans to eventually fire up WANSync SQL and WANSync Server for application servers. He says XOSoft representatives told him Exchange was the hardest part of the suite to implement.

"It's not the easiest thing in the world to learn to run," says Williams, "but it's a whole lot easier than when I used to work with mainframes."

Dave Raffo, Senior Editor, Byte and SwitchOrganizations mentioned in this article: