Preparing for Disaster

The severity of recent storms in the U.S. has sparked many companies to review and revise their disaster recovery plans. Fortunately, Seattle-based Grange Insurance Group ($1.5 billion in assets) has never had to initiate its disaster recovery plans, but the property and casualty insurer knew that its outdated tape backup system, which was putting undue strain on company resources, probably would not provide adequate access to backup data following an earthquake or a volcanic eruption, according to Dale Caldwell, Grange's systems programmer.

It took four hours or more for Grange to copy each day's data to tapes, and the processing would bump up against the insurer's overnight batch processing needed for the next day's business, creating a capacity problem for the carrier's IBM (Armonk, N.Y.) 2003-215 mainframe, which was utilized for both batch processing at night and real-time processing during the day. "There would be many days when we couldn't bring up [the mainframe] on time in the morning because the tape backups had taken so long, which in turn delayed our nightly batch processing run," Caldwell explains.

Furthermore, on average, Grange shipped 40 tapes twice a week to a storage facility 10 miles away, according to Caldwell. "Volume was a real issue," he says. And, given Grange's proximity to the storage facility, the company was concerned that in the event of a regionwide disaster, there was a chance that the backup tape storage site would be inaccessible, Caldwell adds.

So, in 2003, the insurer added a proprietary disaster recovery site in Spokane, Wash., about 230 miles from the corporate headquarters, that utilized Double-Take replication technology from NSI Software (Hoboken, N.J.). While the solution would replicate any newer open storage data (i.e., anything on the insurer's Microsoft -based [Redmond, Wash.], client/server applications), it still didn't address the tape management and storage issue for mainframe data. "Double-Take doesn't speak IBM, so we were still looking for some type of disaster recovery scenario that would eliminate the tapes," Caldwell explains. A secondary issue was the cost to store the tapes - about $1,000 a month, he notes.

So Grange continued to look at replication solutions that would solve the storage challenges and provide more immediate access to backup data. The insurer finally settled on Bus-Tech's (Burlington, Mass.) Mainframe Appliance for Storage (MAS), which was added in December 2004. The appliance translates the mainframe information into a "Windows-like" format that Double-Take then can replicate, Caldwell relates. Installation of the mainframe-attached storage device involved interfacing the Dell (Round Rock, Texas) server that acts as the appliance to the legacy mainframe at the headquarters and the disaster recovery site, a process that took about two hours, he continues.

Data is transmitted to the open disk storage system housed in the MAS appliance via built-in dual gigabit Ethernet ports, and backup data is created in just two hours. Replication occurs almost immediately at the disaster recovery site, according to Caldwell. "The MAS has made a big difference to my work schedule," he says. "It's a very good solution for us."

To test the disaster recovery solution, Caldwell has disconnected the Seattle and Spokane sites and run each for a day on its own, relates Caldwell, who reports that he also tests parts of the system on a quarterly basis, all with satisfactory results.

Case Study Profile

Company

Grange Insurance Group (Seattle; $1.5 billion in assets).

Lines Of Business

Property and casualty insurance.

Vendor/Technology

Mainframe Appliance for Storage (MAS) from Bus-Tech (Burlington, Mass.) and Double-Take data replication technology from NSI Software (Hoboken, N.J.).

Challenge

Eliminate complexity and cost of tape cartridge data storage while providing a viable disaster recovery solution.