Manage Learn to apply best practices and optimize your operations.

Virtualization deployed for improved data center uptime

Virtualization is a technology usually associated with saving space or disaster recovery. But for some organizations, the most pressing reason to deploy it is to ensure data center uptime.

If your plans for server virtualization are focused on cash savings, you aren't alone. However, take a look down the road, and you may see benefits that extend well beyond savings in hardware and operational costs -- you may see improved uptime.

The primary factor leading organizations to deploy server virtualization is reduced costs (operational and capital) as well as improved server resource utilization. "However, some organizations do deploy server virtualization primarily for improved business continuance and as an enabler for disaster recovery," said Mark Bowker, an analyst at The Enterprise Strategy Group Inc. in Milford, Mass.

"The bottom line is that most organizations implement server virtualization for other reasons besides high availability, but it is a welcome benefit that they were unable to obtain in the physical world."

The cost benefits from virtualization can be fairly apparent. When the Florida Department of Education saw that it was time to replace servers three years ago, bureau chief Ted Duncan believed that virtualization had matured to the point where it could help the department save money on hardware. The department virtualized about 30% of its applications, moving the virtualized workload from 60 servers to only six with the help of VMware Inc.'s Infrastructure 3 Enterprise. "We didn't want to spend the kind of money it would have cost us to replace single servers. There was a cost associated with the VMware software, but we did save in that we spent much less on hardware," Duncan said.

However, the education department also found uptime benefits. First, Duncan's team was able to buy all new hardware, which proved more reliable than the aged servers that had been suffering numerous failures. Then, virtualization allowed the department to revamp its disaster recovery strategy by moving virtual machine images to remote servers without worrying about hardware dependencies, which Duncan called "the icing on the cake." In addition, the ability to manage just six servers is "reminiscent of the old mainframe days," in that he doesn't have to monitor 60 independent systems, and it's easier to deploy new applications.

Virtualization allows managers to package in a single file whatever is running within a particular virtual machine, said Laura DiDio, a research fellow at Yankee Group Research Inc. in Boston. "You are encapsulating or isolating applications into virtual containers that you can actually move from one physical server to another. These containers are easy to replicate and make moves and changes," DiDio said.

She added that virtualization does more than support disaster recovery -- it also reduces planned downtime because of fast backups and application testing and deployment, all of which could require scheduled downtime.

We didn't want to spend the kind of money it would have cost us to replace single servers.

Ted Duncan, bureau chief, Florida Department of Education

Bowker added, "The portable nature of virtual machines enables the manual or, ideally, the automated movement of virtual machines between physical servers with zero to limited downtime. If the physical server fails, the virtual machine [and applications] experience a minimal service interruption, or in many cases no application interruption at all."

However, virtualization works only if it's carefully planned out. "A lot of times, people make mistakes because they thought it was easy to put together, but they haven't had enough training," DiDio said. "They make the mistake of not knowing what is on their network, what's working and what isn't. You don't want to take a tired old server and just re-provision it." She advises managers to make sure they buy "rock solid" servers from vendors they can rely on because each physical server will be supporting more applications.

Bowker also emphasized planning. "Poor planning could lead to service degradation. For example if you run many resource-intensive virtual machines on the same physical machine without balancing the workload across the pool of hardware resources, multiple virtual machines could experience performance problems," he said. "It is also important to consider the relationship between applications. For instance, if the [Microsoft] SQL Server fails, but the application remains online, you are still down."

James M. Connolly is a freelance technology and business writer based in Norwood, Mass. Write to him at

Dig Deeper on Enterprise systems management

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.