High-Availability Benefits from Virtualization

Keep It Moving

The Medical College of Georgia relies on a virtualized server architecture to ensure that its more than 5,000 faculty, staff and resident physicians always have access to the services they need.

“Virtualization brings a lot of high-availability options to the table,” says Lawrence Kearney, system support specialist at the Augusta college. “You can find ways to do out-of-band high availability that might otherwise not be available to you.”

Using Citrix Systems Xen or VMware virtualization hosts and clustering software, says Kearney, you can create a script that allows a virtual server to instantly failover to another node in the same cluster, then quickly spin up a new instance of the server – allowing for continuous operation without affecting the user experience.

The key is to determine what high availability means to your business model, prioritize what apps and services really need to be highly available, and how all of that fits into your budget, he adds.

“For example, e-mail, file services and web server access to applications are three things that can never be down,” Kearney says. Apps that are needed by large numbers of users concurrently also get high priority. But some departmental apps that may be used by only 15 or 20 people are generally not made highly available, he says.

At Temple University in Philadelphia, which also includes a teaching hospital with nearly 750 beds, ensuring high availability can literally be a matter of life and death.

“A server failure is not just a case of people being unable to get their e-mail,” says Adam Ferrero, executive director of network services at Temple. “Anything that affects hospital operations needs to be as available as possible. If our network goes down, the mortality rate goes up.”

About 100 of the university's more than 500 servers run VMware, which allows the university to use vMotion to move apps from one virtual machine to another in case of hardware failure. Another key to ensuring uptime is to build in redundancy, says Ferrero.

“We just built a new joint data center for the university health system, and everything in it is completely redundant – all our servers have multiple switches, power supplies and network cards,” he says.

If the university's primary data center fails, a secondary one takes over, he adds. And the university relies on two high-availability Crossbeam X-Series chassis to virtualize its security infrastructure and deliver firewall and intrusion detection software.

Virtualization alone, however, won't guarantee continuous operation. The most reliable solution is to create a virtualized environment using fault-tolerant hardware, to synchronize data processing across multiple virtual machines.

“The lowest level of high-availability requirements can be met by virtual machine software combined with motion technology, but the highest levels of availability cannot be achieved by virtualization because the transition time is too long,” says Kusnetzky.

For environments that cannot tolerate even a few seconds a month downtime – such as electronic funds transfers, where even small delays in transaction time can cost millions – you still need nonstop continuous computers.

“Put in boxes designed for continuous availability, have virtualization software running on them,” Kusnetzky says, “and you'll never see a failure.”