MoneyGram achieves fast, effective disaster recovery

When a few minutes of downtime can mean thousands of lost transactions and millions in lost revenue, you’d better have a disaster recovery plan. MoneyGram, a global payment services company, had a plan alright — except that it could guarantee only that its system could be restored within 24 hours. Upper management realized the risk was way too high.

“It was clear that we needed a more effective disaster recovery plan and strategy,” says Tom Becchetti, senior capacity planner at MoneyGram. “We needed to eliminate outages, increase uptime, and provide more availability for our products and services.” In fact, management declared that downtime in the event of disaster would have to be reduced to zero.

To meet this ambitious goal, Becchetti began by using VMware to virtualize server resources across four geographically dispersed datacenters. His team then built a 160-port SAN using two EMC Symetrix DMX storage arrays, a Clariion disk-to-disk backup system, and McData Intrepid 6140 Directors with FICON (Fiber Connectivity) interfaces. Finally, a disaster recovery site was created to handle the entire load of the main datacenter, effectively achieving the zero-downtime goal. In the bargain, MoneyGram is saving a significant amount of cash per year: $400,000 in network connectivity; $120,000 by switching to FICON interfaces; and as much as $500,000 by switching from Unix to Linux.