IBM Mirrors Shark Over FC

IBM Corp.'s (NYSE: IBM) Shark will be swimming faster -- and farther -- soon. Big Blue today announced several enhancements to its TotalStorage Enterprise Storage Server (a.k.a. Shark), including Fibre Channel support for its Peer-to-Peer Remote Copy (PPRC) disaster recovery software (see IBM Boosts Shark's DR Features).

By replacing its Escon technology with Fibre Channel, IBM claims the new Shark will improve data recovery by transferring data up to eight times faster and reducing channels by a factor of four. IBM said it would introduce the PPRC-over-FC capabilities when it last refreshed its array-based software for the Shark this May (see IBM Pushes Shark's Copy Buttons).

IBM also announced a new API for Copy Services -- based on the Storage Networking Industry Association (SNIA)'s Storage Management Initiative Specification (SMI-S), which gives third-party management tools access to its copy services -- and a Turbo II processor option that the company claims will offer a performance increase of up to 30 percent over the base ESS 800 model. The company also has a new configuration option called Arrays Across Loops, which it says can nearly double the performance for sequential applications on a single array. IBM says each of the new enhancements will be available Nov. 21.

While IBM was merely playing catch-up with competitors EMC Corp. and Hitachi Data Systems (HDS) with Fibre Channel over PPRC, it claims its new enhancements give it a genuine performance advantage.

"We think our collapse rate and performance is significantly better than our competitors'," says Jim Tuckwell, marketing manager for Shark. "We do not see a situation where a customer will need more than two Fibre Channel links. What we've seen from our competitors is, they offer a minimum of four Fibre Channel links."Adds Tuckwell: "We think we've got a performance advantage that equates to fewer channels, and with fewer channels comes a cost savings."

Analyst Mike Fisch of The Clipper Group Inc. says he's not sure about the performance advantage, but he agrees that the enhancements have improved the Shark product line.

"I dont think IBM is pulling ahead, but this is certainly better than Escon," he says. "You have a three-way horse race among high-end storage vendors with EMC’s Symmetrix, Hitatchi’s Lightning, and IBM’s Shark. They’re all constantly jostling, and it’s IBM’s turn to jostle in that group."

According to IBM, the Shark’s new long-distance data mirroring capabilities will improve disaster recovery and reduce costs. To show the Shark’s extended reach, the company today demonstrated a data mirroring implementation spanning 1,200 miles from to Minneapolis to Lumberton, N.J. [Ed. note: Yes, Lumberton, N.J., world-renowned hotbed of data replication!] The demonstration was made in conjunction with CNT (Nasdaq: CMNT) (see CNT Extends IBM Data 1,200 Miles).

The Shark’s improved disaster recovery manageability will also allow changes to be tracked to data on backup volumes while the primary storage is unavailable. These failsafe measures help to greatly reduce the time needed to resynchronize PPRC volumes after switching from the primary to the remote site, IBM says.“This will have interest for customers who want to do server clustering and failover between remote sites,” Fisch says. “This is incremental improvement for Shark on a number of fronts, but add it all up, and it’s a significant announcement for their high-end storage product.”

Tuckwell says the new PPRC version and the Turbo II processor will be priced the same as their predecessors. Shark systems range from $300,000 to more than $1 million, depending on configuration.

— Dave Raffo, Senior Editor, Byte and Switch