Article

Virtualization increasingly used to abet disaster recovery

Linda Tucci, Executive Editor
George White, CIO for the Pennsylvania Office of Attorney General, moved the office's mission-critical IT into the virtual world, switching out 150 old, standalone servers for 50 Dell Inc. blade servers and VMware Inc. virtualization software.

    Requires Free Membership to View

My CEO says, 'We're not going to spend that; go find another solution.
Mike Carvalho
CTORadiator Express Warehouse Inc.
"It was scary in the sense it was new technology for us," White said. The results were so positive, in cost savings and performance, that he plans to implement virtualization at the office's remote disaster recovery site starting this spring. "We're taking virtualization in a building-block manner."

The CIO from Pennsylvania is not alone.

Six months ago, when the subject of virtualization raised its seductive head, the talk among midmarket CIOs was all about shedding pounds, said analyst Jim Browning, who covers the midmarket at Gartner Inc. in Stamford, Conn.

"They took their 30 servers and cut them down to 20, saved $50,000 and felt good about themselves," Browning said. "Then they quickly realize virtualization can be used for other things."

At a recent gathering of some 350 midmarket CIOs in Palm Springs, Calif., Browning noted that virtualization is increasingly being used as a vehicle for affordable disaster recovery -- and the troops are feeling even better about themselves.

"Midmarket CIOs have been getting a lot of pressure for years for improving disaster recovery," Browning said. "They are taking some of those servers they shut down, virtualizing them and putting them in some remote location and actually replicating applications. Email and ERP are the applications I hear about most often."

Disaster recovery in the physical world is hard for midmarket companies. A company looking to set up a second recovery site needs to duplicate its hardware. That is certainly expensive, especially if you live by the rule of one application per server. Maintaining a second site more than doubles the trouble. Every patch job at the primary site requires a patch at the secondary one. The third-party providers that can make disaster recovery less painful for big enterprises are too expensive for many midmarket companies.

As the cost of IP networks, ISCI and disk-to-disk backups have become more affordable, disaster recovery-by-virtualization is growing in popularity, said Browning and others. "VMware is a rock star to them right now." Browning said, referring to the Palo Alto, Calif.-based market leader. "In fact, virtualization and VMware are interchangeable. They're using VMware as a verb."

Because of budget constraints, White wasn't able to duplicate the server environment, traditionally or virtually, at the secondary site. But the foundation is in place. "We implemented a dual storage area network at our primary site in Harrisburg, the capital, and our discovery recovery site in State College, Pa. From a storage standpoint, the primary site is mirrored up to the disaster recovery site at regular intervals," White said.

In the meantime, White has implemented clustering within the on-site environment, as well as dynamic failover capabilities so in the event of a problem, the virtual machines will automatically revert to an area that is working properly. "Live migration capabilities mean we can manually move environments around to do routine maintenance without disrupting users," White said. "We're using a combination of technologies locally as a stopgap solution until we can have a full-scale virtualized environment deployed at our DR site."

Mike Carvalho's embrace of virtualization required a big shove from above. The CTO of Radiator Express Warehouse Inc. (1-800-Radiator), Carvalho has been working in IT since 1981, when he joined the U.S. Air Force. "I can be stuck in my old ways," he said. So when a buddy took him out to lunch to sing the praises of virtualization, "I said, 'Man, you're not going to sell me on that,'" Carvalho recalled. "I bit into the model of one physical sever for one application."

Problem was, his company was adding franchisees at a rate of two a week and the data center was running out of power and flat out of space. 1-800-Radiator is a privately held automotive parts distributor in Benecia, Calif., that does approximately $80 million in revenue. Part of the franchisee deal is that the company's headquarters manages all the IT for its 200-some locations. "I analyzed what it would cost to keep up with the pace of growth, given the number of servers I was adding, and put together a quote. My CEO says, 'We're not going to spend that; go find another solution.'"

It took three weeks to virtualize 31 servers, using products from VMware. Power consumption was cut by 25%. Then another light bulb went off. Like many midmarket companies, 1-800-Radiator didn't have much in the way of a disaster recovery strategy.

"We had a couple of MetaFrame servers sitting in a colo someplace and we were going to use a DOS application for disaster recovery," Carvalho said.

Instead, he took some of the money he saved on physical servers and bought three more VMware servers. Only critical business functions -- load balancers, Web servers, database servers and so on -- are replicated, so "I was able to duplicate everything I needed in two racks," he said. "We staged it, we built it and started testing it. Today my disaster recovery site is 24/7 live." The remote location is unmanned, another attractive feature of the virtualized environment. Carvalho uses the VMware management tool to tend the servers and the operating systems, "and I'm doing it all over a WAN," he said.

More on disaster recovery
Disaster recovery planning for CIOs

Power outage brightens Las Vegas utility's outlook on virtualization
VMWare's High Availability is enabled on all servers at both locations. The VMotion tool, which automatically moves a virtual machine from one server to another as needed, is deployed judiciously on noncritical servers, because "I don't want the risk of any corruption," Carvalho said.

What he likes best about the disaster recovery solution, he said, is that it is "constantly being tested." His staff monitors the networks using the latest Pro version of IPSwitcher from Softmate Corp. The disaster recovery servers are checked every 60 seconds. "My guy makes sure the Web sites come up, and I know the software code is current. I can go home and sleep, is what it comes down to. I am not worried about disaster recovery."

A good night's sleep may be the result, but cost savings was the primary driver.

"It took the situation I described -- a lack of resources -- to actually see it," Carvalho said. "The company wants to grow. My job is to make sure it can grow the way it wants to. We can't say, 'OK we're not going to grow this week, because Mike has decided we don't have the resources.' They're going to find somebody else who does.'"

Let us know what you think about the story; email: Linda Tucci, Senior News Writer


There are Comments. Add yours.

 
TIP: Want to include a code block in your comment? Use <pre> or <code> tags around the desired text. Ex: <code>insert code</code>

REGISTER or login:

Forgot Password?
By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy
Sort by: OldestNewest

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to: