No Rest for Summit’s Retired Servers

Virtualization projects have won lots of praise as a wonder drug to fight energy headaches by reducing servers. But it’s not magic-the left-over servers don’t disappear into thin air. One possible solution can be found at Summit National Bank, which is using the servers as part of a new disaster recovery site.

The $75-million asset bank, which has four branches in Wyoming and Montana, started its virtualization project in January, buying a HP ProLiant ML370 G6 with 64 gigabytes of RAM, 2.5 terabytes of storage and a two-quad processor, on which it virtualized nine servers using Microsoft’s Hyper-V software. The project resulted from a desire to add the power of four servers to its five existing IBM xSeries 306m physical servers to accommodate new products and services.

Jonathan Tinsley, Summit’s CISO, says the bank decided against a cloud service for remote backup out of concerns over control-which led the bank down the path to using its old servers as part of disaster recovery. The bank also had to select a provider. “We evaluated Symantec Backup Exec and AppAssure Replay,” says Tinsley. “Replay cost around $4,000. Symantec wasn’t any cheaper, and Replay seemed to handle restores a lot faster.”

Summit used one of the IBM servers as a local back-up server running Microsoft Backup, and another as a remote back-up server at another branch. The ProLiant runs a Replay agent program, with the two back-up servers running the core Replay back-up application.

The Web link for the remote back-up system was scheduled to go live by the end of August, Tinsley says.

Summit put the memory and hard drives from the remaining three decommissioned servers into the remote server.

AppAssure CEO Najaf Hussain says Replay is able to minimize the storage space and other infrastructure needed for backing up mission-critical applications and data because of its use of deduplication technology. “Deduplication means you only back up a file that has actually changed,” he says. “Older backup systems make a new backup copy every day regardless of whether the data has changed.”

Replay uses an image-based approach for backup, says Steve Scully, research manager for continuity, disaster recovery and storage orchestration at IDC. “It takes a snapshot at a moment in time, which is very viable in a virtualization environment. Whether it’s worth re-using servers decommissioned due to virtualization depends on their age and capability. If the servers have some life left in them, it makes sense to use them for backup. After all, the likelihood you will have a disaster and need to use the backup servers is low.”

But the strategy does have a defined shelf life, considering the fact that the legacy servers are legacy servers for a reason.

“If a bank wants to use old servers for backup, it should realize that they will probably only be good for two to three years,” Marc DeCastro, research manager, consumer banking and community banking at IDC Financial Insights, says. “Then the bank will need to replace them.”