Panasonic Joins the SAN Age

When Panasonic, the main North American subsidiary of Matsushita Electric Industrial Co. Ltd. (NYSE: MC; Tokyo: 6752), picked McData Corp. (Nasdaq: MCDTA) for its first SAN last year, the main impetus was to streamline the company's storage and make it more manageable.

Despite being a hub for application access from more than 100 North and South American company locations, Panasonic's Secaucus, N.J., production center, was using an old-fashioned Serial Systems Architecture (SSA) approach to storage networking from IBM Corp. (NYSE: IBM). Each IBM server had its own storage disks, and when something went wrong, each server in the network had to be checked. "We only used generic AIX management," says Craig O'Connor, Panasonic's systems program project leader.

SSA was IBM's early alternative to Fibre Channel, but by most accounts its day is done. "IBM is the only one who ever implemented SSA, and it's pass," says consultant Arun Taneja of the Taneja Group.

O'Connor says even IBM agreed that a SAN was the way to go. So two years ago, he and the Panasonic team chose switches from McData Corp. (Nasdaq: MCDTA) to create a Fibre Channel fabric over an 18-month period.

O'Connor says his team's choice of FC directors was between McData and Inrange, now part of Computer Network Technology Corp. (CNT) (Nasdaq: CMNT). (See 2003 Top Ten: Mergers & Acquisitions.) McData won out, he says, because its SANavigator management software impressed the team more with its provisioning and monitoring features (see McData Helps Panasonic).Panasonic started by implementing a single FC fabric using McData Intrepid 6140 Directors and Sphereon 4500 Fabric Switches at the company facility. A second fabric was installed for redundancy a bit further into the project. Panasonic now has two Intrepids and 10 Sphereons in Secaucus. The SAN supports a total of 312 ports, all in Secaucus.

IBM is still the company's storage vendor. Indeed, the number of IBM servers at Panasonic has actually increased from about 50 to 90, as data centers in Mexico, Brazil, and elsewhere have been consolidated at Secaucus. Now, however, thanks to the SAN, two IBM Shark Model 800 arrays have replaced 10 racks of SSA disks that used to be attached to individual servers. Storage can also be provisioned and managed centrally, instead of at individual servers.

It's the easier management and administration that's made a dramatic difference, O'Connor says. Scaleability has improved. Since there's better visibility into actual utilization, it's possible to be more efficient in how storage is used. Two administrators have replaced five. What's more, each administrator manages more -- about 38 Tbytes of data, compared with the 5 Tbytes each managed in the old setup. In total, O'Connor says, the SAN's manageability has reduced labor troubleshooting by 25 percent annually.

O'Connor won't say how much the whole thing cost -- at least not right now. But one thing is clear: He's pleased with the money he's saved.

— Mary Jander, Site Editor, Byte and Switch0