Many large companies try to maintain hot sites that are in lockstep with the production environment, but this disaster recovery plan isn’t always realistic.
Configurations drift, or IT staff simply don’t have the time — or the budget — to mirror every aspect of their production environment. That’s where virtualization comes in.
Applications that may not have made it on the mission-critical DR list can now be put on a shared piece of hardware: a virtual server in the data center or hot site. The costs of maintaining hardware for both mission-critical and not-so-critical applications can drop considerably in this scenario.
The other day, Nelson Ruest, a principal with consulting firm Resolutions Enterprises Ltd., in Victoria, British Columbia, was telling me that one of his enterprise clients is projecting 60% to 70% savings, per year, across its infrastructure for disaster recovery. The company moved from mirroring the production environment — taking up three floors to do so in its backup data center — to using less floor space and hardware with a virtualized DR plan. It replaced many physical servers with virtual servers in its production environment as well.
Independent Bank, out of Iona, Mich., saved $1 million in hardware by replacing underutilized hardware with virtual servers, as part of a desktop-to-data-center virtualization DR strategy.
The shift to virtual desktops and servers allowed the bank to eliminate configuration drift between its hot site and production environment as well, according to its CIO, Pete Graves.
Still, enterprises are hesitant to use virtual disaster recovery for mission-critical applications, and are definitely not throwing out tape backups any time soon, according to John Humphreys, senior director of product marketing for the virtualization and management division of Citrix Systems Inc.
Humphreys is seeing the spread of virtualization for DR take the same path that the technology did on the server front: around the edges of the enterprise, starting with non-mission-critical applications that haven’t made it into the “critical” DR budget.
Others believe that virtualization DR will continue to evolve as server virtualization does. As the staff at SearchServerVirtualization.com point out in an article on predictions for 2010, there is still the potential for one virtual server to take down hundreds of other virtual machines with it.
So, it would seem that enterprises are testing out and moving forward with virtualization disaster recovery, but with a note of caution.
Tell us what your disaster recovery plans are and if virtualization will be part of them. Email firstname.lastname@example.org.