Q. Why does virtualization help disaster recovery?

A. In a standard disaster recovery environment, you have a primary data center and another location, with copies of your physical hardware, for site disaster recovery. If the hardware isn't the same at the two locations, you may have driver incompatibilities and problems restoring your backups.

With virtualization, your "hardware" is all virtual and completely separated from the actual, physical hardware in the host server. This separation means it's much easier to take a virtual machine (VM) and restore it to different hardware, as long as both servers are running the same virtual platform (hypervisor) and have comparable resources.

Another factor with virtualization is that you simply have fewer machines to worry about. If you go from 50 physical machines to 10 virtual servers, you've reduced the number of restorations you have to do by 80 percent.

Most virtual platforms, including Hyper-V, support the Microsoft VSS backup approach. This approach means that you only need to back up the host using VSS, and all the host's VMs are backed up as well. Your VMs are notified of the VSS backup and ensure the data on disk is in a consistent state. See this FAQ for more on VSS backups of VMs.

You should make sure your plan for recovery is sound when using virtualization. You should be either performing VSS backups or using a platform like System Center Data Protection Manager. You could also export your VMs, which would let you get configuration information about them. You could then import the VMs at your disaster recovery location using this information. If you plan to do this, make sure you frequently back up your VHDs or have them replicated between locations using a SAN. If you have a replicating SAN available, you may be able to do a multi-site cluster instead, and remove the need to have to export/import the configurations. See this FAQ for more about restoring a VM from a VHD.

Microsoft has provided a good TechNet article that gives more information about Hyper-V high availability scenarios.