MidAmerica Bank

Illinois-based MidAmerica Bank has become a poster child for business continuity, thanks to a multivendor effort that replicated the company's data center to ensure faster recovery.

In the recently completed project, MidAmerica replaced its "cold site" disaster recovery plan with an in-house "business continuity site" that features data replication and mirroring across the same kind of tiered storage devices it has in its data center. The data center is linked to the remote site over a 1-Gbit/s IP service from the local exchange carrier SBC Communications Inc. (NYSE: SBC).

MidAmerica's project is interesting on a couple of counts. First, it shows how dramatically a company can improve its disaster recovery by implementing disk-based replication techniques, as opposed to relying on tape backup and off-site recovery.

According to Paul Stonchus, first VP and data center manager at the bank, the recovery time with tape vaulting was a matter of days, for a single application. Now it's down to less than an hour for 25 applications -- clearly a must in this 21st century of terrorist attacks and grid power outages. MidAmerica, the $9 billion subsidiary of MAF Bancorp Inc., also had its own growth to consider, since it's burgeoned from 16 remote branch offices to more than 60 over the last eight years.

MidAmerica enlisted its storage vendor, EMC Corp. (NYSE: EMC), for design help. The vendor recommended that the bank pick a site "within driving distance" of its headquarters data center in order to set up the business continuity site.The BC site is based on a SAN comprising four UltraNet Edge Storage Routers from Computer Network Technology Corp. (CNT) (Nasdaq: CMNT) and an unspecified number of FC switches from Brocade Communications Systems Inc. (Nasdaq: BRCD).

At the new BC site, mission-critical banking data is stored on the most expensive Symmetrix arrays attached to mainframes from Unisys Corp. (NYSE: UIS), just as it is in the bank's data center. The data is mirrored and backed up with EMC software. Less critical apps like loan tracking, ATM authorization, and email are assigned to CLARiion CX Fibre Channel and ATA arrays, with accompanying software for mirroring and backup.

Out at the branch offices, MidAmerica uses Double-Take software from NSI Software Inc. to consolidate backup data for delivery to the data center and BC sites.

This tiered approach, which EMC and MidAmerica cite as an example of ILM (information lifecycle management), is a money-saver, Stonchus says. "[W]e use SRDF/S software to synchronously mirror our mission-critical data that is stored on Symmetrix and with no impact on our applications," he states in a document prepared by EMC. "We accomplish the same with MirrorView software for our mid-tier application data stored on CLARiiON, but at a lower price point. In addition, by using SRDF/A to mirror our test data asynchronously, which can tolerate some level of exposure, we're able to balance recovery performance and bandwidth costs."

But just how much money MidAmerica is saving isn't clear. The company says it won't comment on its network except what it's approved for its suppliers to broadcast. Indeed, the MidAmerica situation has been widely hyped by the vendors involved, particularly EMC.And none of the vendors is willing to talk money. Still, considering the amount of equipment involved, it's a safe bet the project racked up a multimillion-dollar pricetag before it was finished. That said, however, the off-site services of disaster recovery providers like ManagedStorage International Inc. (MSI) can cost thousands per month per application, so it's likely money well spent compared with its previous tack.

Mary Jander, Site Editor, Byte and Switch