How virtualization eases disaster recovery

Contact Us

Contact a Microsoft Representative

Your satisfaction Matters!Let us know your thoughts about your Microsoft experience.




Related Links

Virtualization: Business Continuity/Disaster Recovery

Virtualization Overview

Virtual Machine Manager Overview

Disaster recovery is a daunting challenge for any business. Fortunately, a new technology you may already be deploying can help.

Virtualization, a versatile technology with multiple ramifications for companies, reduces the price and complexity of business continuity. Essentially, by using virtualization technology, you can encapsulate an operating system, an application, and its data into the equivalent of an application running on top of an operating system; the operating system, in this case, is the virtualization software. You can then easily transmit this encapsulated application to an offsite location—just as you would transmit a data file—and make it available on a remote machine for employees to access. This capability alone can cut downtime in the event of system failure from days to hours, or less.

To make virtualization work for your disaster recovery strategy, consider these tips.

Calculate your performance needs. "Before virtualization, you had to maintain a one-to-one relationship between your production servers and your standby servers," says Jack Zubarev, COO of SWsoft, a Virginia-based Microsoft Gold Certified Partner specializing in virtualization. "That's no longer the case. You can replicate five production servers on a single server running multiple instances of a virtual operating system.

"Just mapping five machines onto one, however, doesn't mean an automatic 80-percent reduction in costs," Zubarev says. "But it's true that having one server dedicated to disaster recovery is less expensive than having five. You must still consider performance: Running multiple operating systems, even in virtual mode, requires sufficient memory and disk space on the backup server."


*Before virtualization, you had to maintain a one-to-one relationship between your production servers and your standby servers. That’s no longer the case. You can replicate five production servers on a single server running multiple instances of a virtual operating system.*
Jack Zubarev
COO, SWsoft

Consider application dependencies. Most virtualization software, including Microsoft System Center Virtual Machine Manager 2007, offers tools to determine which applications and services to restore in which order. This is important not only because of dependencies within applications, but also because IT needs to understand, from a business standpoint, which applications should have top priority. "If Microsoft Exchange Server is the most important application, you have to make sure the domain controllers are restored first, because they contain the [Windows Server 2003] Active Directory information," explains Martin Thieret, vice president for Enterprise Infrastructure Services at Colchester, Vermont–based Competitive Computing, a Microsoft Gold Certified Partner specializing in virtualization. At the same time, use the virtualization software to make sure that you transfer and set up the operating system, application, and data in that order.

Always test the configuration. Using virtualization within your disaster recovery strategy gives you a lot of flexibility. Because bandwidth costs have decreased significantly over the past few years, it's now more cost-effective to set up active-active connections or active-passive connections. In an active-active connection, one site—say, another branch office—acts as the backup for another and vice versa. In an active-passive connection, one site—such as a service provider for disaster recovery—remains dormant until it is needed. In the latter configuration, you can contract with third-party hosting services to set up the virtual machines. Whichever configuration you choose, test your ability to recover. That's good advice with any disaster recovery strategy, but it's especially important when you're adding new technology.

Even the convenience of virtualization doesn't abrogate the need to have a clear set of procedures outlining roles and responsibilities ahead of any business continuity or disaster issues. As Zubarev notes, "If you don't have a plan, it doesn't matter how good your underlying infrastructure is."


Howard Baldwin

Silicon Valley-based freelancer Howard Baldwin writes regularly for the Microsoft Midsize Business Center. His work has also appeared on AllBusiness.com and in CIO.



Was this information useful?