Mainframe > Case Studies > Banking/Finance

A Sound Investment

Citigroup tames its backup environment with dedicated mainframes

“IBM has some strong algorithms built into the machines, as well as crypto engines and certificate management.”

Up Close

Customer: Citigroup Inc. Headquarters: New York City Business: Global provider of financial products and services for consumers, corporations, governments and institutions Challenge: Improving data-backup processing Solution: Using four IBM System z10 Business Class machines as dedicated backup servers and INNOVATION Data Processing’s FDR/UPSTREAM and FDRSOS to automate backups Hardware: Four IBM System z10 Business Class servers Software: INNOVATION Data Processing’s FDR/UPSTREAM and FDRSOS

Whether for internal or external reasons, most organizations require data be backed up and available for disaster-recovery or business-continuity purposes. In particular, financial institutions are often held to a higher standard because of the sensitive nature of the data they handle.

Although most customers “don’t see the detailed processing that happens in the background, the fact that we can tell them that their data is being managed with the highest levels of integrity, availability and security is a big plus,” says Martin Kennedy, Citigroup Inc.’s managing director of enterprise-systems infrastructure.

That’s why Citi created a backup environment many would consider best-in-class, using four IBM System z10* Business Class (BC) machines as dedicated mechanisms for backing up the company’s myriad servers and systems. Combined with INNOVATION Data Processing’s FDR/UPSTREAM and UPSTREAM/SOS, their design not only ensures proper backups, but also optimizes CPU usage and network bandwidth.

“The traditional backup model limits backups to an off-hour window because of the high CPU and bandwidth requirements,” says Kennedy. “With our model—with the data moving through a dedicated SAN [storage area network] fabric and not over TCP/IP—you can pretty much schedule backups whenever you want without impacting production applications and workflow.”

Tweet