The replication project sprang out of a data backup project to protect the Northwest Regional Data Center's (NWRDC) IBM z/OS mainframe, which it chose to do using Luminex Software Inc.'s Channel Gateway virtual tape system (VTS). The Luminex VTS converts mainframe FICON data into Fibre Channel format for storage on Nexsan's open-systems SATABeast array and replication to Atlanta using the Sanrad switch.
NWRDC software support manager Mikal Haney said the Nexsan/Sanrad combination cost around $1.2 million, while he would have spent at least $2 million on other options he considered.
EMC, IBM and Cisco in running
Haney looked at three other options when NWRDC was planning the project, including two array-based projects. He considered a bid from EMC Corp. for Symmetrix DMX-4 disk arrays and SRDF replication software at both ends of the wire. He also had a bid from IBM for additional capacity on the DS8100, which NWRDC uses for on-site primary mainframe storage and Global Data Replication (GDR) to an automated tape library (ATL) at the other end of the wire. "It didn't make sense to me to take DASD [direct access storage device] and replicate it to an ATL in a cage in Atlanta when tape wasn't something we wanted to use in the first place," Haney said. "That was the only solution IBM quoted." The other option was a Cisco Systems switch that he said cost three times the V-Switch from Sanrad.
Weigh scalability considerations
Haney said cost was the most important consideration in implementing the VTS system with Nexsan and Sanrad. The bid for the Sanrad system was submitted, and the equipment eventually deployed, by local VAR/storage integrator LH Computer Services in Coral Springs, Fla.
"We're working to tie more schools into the replication system now," said Don Ferreri, founder and partner at LH Computer. NWRDC has 130 TB of space on the Nexsan arrays at each end of the wire, with about 75 TB used. Haney estimated it's sending 3.5 TB of data per day over the wire using six of the Sanrad switches.
Haney said he hadn't seen any drawbacks yet in the deployment. However, Forrester Research senior analyst Stephanie Balaouras cautioned that organizations sending a greater amount of data over the wire should be careful when evaluating an appliance-based approach such as Sanrad's. "Appliance-based approaches to replication aren't bad; they're a bit of the best of both worlds; storage- and server-agnostic. "The only caution is scale," Balaouras said. "I think it's a good solution for midsize environments, but not to replicate hundreds of terabytes."