Virtualization Makes Disaster Recovery Cheaper


Want the latest storage insights?

Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure

No one has to convince Chris Steffen about the benefits of server virtualization for disaster recovery. The Microsoft (NASDAQ: MSFT) Hyper-V user has seen those benefits firsthand.

Steffen, principal technical architect at Kroll Factual Data in Loveland, Colo., said virtualization has led to a big improvement in disaster recovery preparedness at the provider of business information to mortgage and consumer lenders, property management firms, and other businesses.

"Virtualization has made disaster recovery extremely easy," he said. "It literally is a matter of copying a file to a computer and getting it to run."

Through virtualization, the company has been able to reduce its server population by 40 percent and eliminate lots of administrative headaches. "I can't even begin to explain how much easier, fault-tolerant and just simpler to manage it is compared to the solution we had before virtualization," he said.

In a perfect world, the hardware configuration for disaster recovery should mirror the configuration of the primary system. "That's a very expensive proposition, and in the real world, most people are forced to prioritize," noted Ian Robinson, PowerVM offering manager for IBM's (NYSE: IBM) Systems and Technology Group.

With virtualization from the likes of Microsoft and VMware (NYSE: VMW), organizations desiring a full mirror of their primary systems can do so without duplicating the hardware for those systems, Robinson said. "Rather than having a second replica system, you can have an image of the primary system that runs on a virtual machine, or logical partition," he said. "That dramatically saves on the cost."

For example, Patrick O’Rourke, group product manager for Windows infrastructure at Microsoft, said he's seen customers running as many as 15 applications in a virtual environment on a single server. "Five years ago, each of those applications would have to run on a single server," he said.

Virtual Storage Savings

Similar savings can be achieved on the storage side of the disaster recovery equation. There disaster recovery could double storage costs by requiring that hardware doppelgangers of the primary storage configuration be located at the disaster recovery site. Virtualization allows configurations to vary between the primary and the DR site, according to Chris Saul, marketing manager for IBM's SAN volume controller.

"With virtualization, we can have a different physical configuration at the recovery location, but we could still build the same virtual storage configuration on those physical resources," Saul explained.

That opens the door to savings at the DR end of things for companies willing to take a performance hit when disasters occur, he reasoned. Customers are increasingly prepared to accept lower performance for a while following a disaster until they can get their primary storage systems back online, he said. By allowing customers to use a lower-cost, lower-performance storage configuration for DR, virtualization can reduce the overall cost of building a disaster recovery infrastructure.

For companies unwilling to tolerate a performance hit, joining a disaster recovery pool may be an attractive alternative. SunGard, of Wayne, Pa., offers such a scheme.

"We've leveraged virtualization in a shared pool model to take advantage of the good things about virtualization, but in a way that helps minimize the cost of a secondary data center," said Don Norbeck, SunGard's director of product strategy and virtualization technology officer.

"Our secondary site is set up in such a way that each of the companies replicating into it only take up a portion of the pool," he said. "At their time of need, we can very rapidly — in most cases in as little as 30 minutes — extend their footprint to mirror what they would have in production to give them a full production capability. It gives the benefits of one-to-one replication at the cost of many to one."

Page 2: Locating DR sites and backing up VMs

Submit a Comment


People are discussing this article with 0 comment(s)