SunGard Offers Virtual DR Serivice

Ever since its 2001 acquisition of Comdisco Availability Services SunGard has been the dominant player in the disaster recovery business offering a wide array of services, all based on physical hardware. Now, through partnerships wit VMWare and Double-Take, they're entering the twenty first century with a disaster recovery solution using virtual servers at the DR site to receive host based replication data.Server virtualization has revolutionized the economics of the DR site by replacing the rack full of server that are doing nothing but receiving replication data and waiting for the day they get sent into the front lines with virtual servers. The one limit on how much you can save this way is that the virtual server hosts have to be beefy enough to handle the real application loads. Organizations with recovery point objectives (RPOs) of at least several hours, for some applications, can get a bit of a boost by combining VMWare Infrastructure with DR provider's "Shared Server" offerings that make hardware available when the customer's declared a disaster and Vmotioning machines to the shared servers when they're needed.

SunGard's taken this one step further by using a dedicated set of virtual servers as data catchers and moving those virtual machines to another pool of systems when a disaster strikes. This allows them to align the systems with the different requirements of data catching and application serving. Because the WAN bandwidth available natually limits the replication traffic data catchers can use high capacity SATA drives while the application servers run off high performance SAS or FC disk.

They've also taken a soup to nuts approach so the $1000-1200 a month a customer pays for the service includes everything but WAN bandwidth. No worries about power, cooling software maintenance Etc.

Most impressively their promising the while thing can be off the ground in 30 days. Most SME organizations would still be ordering hardware in 30 days not replicating their valuable data.