The Project\u2002::\u2002Implement a hypervisor-based replication system to provide more streamlined business continuity and disaster recovery for Woodforest National Bank\u2019s virtualized data center.\n \n The Business Case\u2002::\u2002Ever since Hurricane Ike blew through town three years ago, business continuity and disaster recovery have become more than an academic exercise for Woodforest National Bank, located just north of Houston. Now each June, all production applications are preemptively failed over to a secondary data center in Austin to prevent storm-related outages, then returned to the primary site by late October. But the company\u2019s traditional, array-based replication process, designed to work in a physical server environment by backing up applications at the storage layer, was clunky when dealing with bank\u2019s 95-percent-virtualized primary data center. The cost savings and efficiency that server virtualization provided to the bank, which manages 750 branches and $3.3 billion in assets, were being eroded by the annual migration. So Woodforest CTO Richard Ferrara signed on as one of the first customers of a hypervisor-based backup and recovery option providing replication further up in the stack.\n \n First Steps\u2002::\u2002Zerto, maker of the new hypervisor-based replication product, approached Ferrara last year. \u201cThey knew I would be willing to give it a shot because I like to be on the edge of things,\u201d he says. Ferrara and his team toyed with the pre-alpha release in late 2010 and implemented the first official release in the bank\u2019s infrastructure lab earlier this year.\n \n With its previous SAN-based replication, recovery activities were complex and manual. In the bank\u2019s data center, multiple virtual machines might reside on a single array volume; array-based backup required Ferrara to replicate the entire volume even if only one of the virtual machines needed backup, resulting in underused storage and sprawl\u2014the very problems virtualization was supposed to solve. In addition, his network managers had to reconfigure the storage drives pre-replication. \u201cZerto does it at the virtual server layer,\u201d says Ferrara. \u201cIt enables us to be more efficient with storage, replicating only the virtual machines we wish to failover.\u201d\n \n Virtual backup and recovery is faster, Ferrara says. When Ferrara used Zerto to move everything back to home base in October, the migration took less than two hours, compared to eight hours the old way. And the 15- to 30-minute downtimes the bank had come to expect as applications moved back and forth were virtually eliminated.\n \n Ferrara sold the system to bank executives on significant cost savings. Array-based backup licenses were priced based on the number of virtual machines Woodforest replicated. With Zerto, Ferrara pays for total terabytes of backup, resulting in a 30 percent savings.\n \n What to Watch Out For\u2002::\u2002As an inaugural customer, Ferrara had to help Zerto work through some kinks, leading to the creation of more user-friendly screens and a more efficient journaling processes. But all in all, \u201cwe haven\u2019t seen any downside\u201d to early adoption he says.\n \n However, virtual disaster recovery and business continuity is not for everyone. \u201cIf you are a looking for physical protection for your physical hosts, Zerto\u2019s technology may not be the right solution for you,\u201d says Ferrara.