Hitachi Data Systems (HDS) Corp. today unveiled a new software capability it will ship later this year called High Availability Manager (HAM). The software automates failover between pairs of Hitachi Data Systems' Universal Storage Platform V (USP V) controllers, and customers say it will make data migrations and upgrades easier.
HAM would be deployed with two USP V controllers attached to external storage. The front-end USP V controllers can be populated with internal storage or used as gateways to external disk. If the front-end USP Vs are populated, Hitachi Data Systems' TrueCopy replication mirrors data between the systems.
The High Availability Manager software then creates a quorum disk for system-state information using external storage. If one controller fails or needs to be taken down for maintenance, HAM coordinates the redirection of application data requests to the surviving controller while applications retain their access to data without interruption.
Hitachi Data Systems has previously been able to nondisruptively migrate data between external arrays, and has offered local and remote synchronous replication software for disaster recovery (DR) purposes, but redirecting servers between front-end USP V controllers has always been a manual process.
Christopher Crowhurst, vice president of strategic technology for the Professional Division of Thomson Reuters, said he previously achieved redundancy in Hitachi Data Systems environments by setting up multiple paths from each server to ASICs within the same array. All of the data still remained under the same controller.
"I haven't experienced a situation where we've had availability problems [with the single-controller USP V]," he said. "We plan to use this more as a nondisruptive way to swap out infrastructure."
Claus Mikkelsen, HDS' chief technology officer of storage architectures, said Hitachi Data Systems has focused on improving the availability of the single controller, but found some customers were still nervous about putting all of their data eggs behind one Hitachi Data Systems controller basket. "This eliminates the virtualization controller as a single point of access failure," he said.
A bigger refresh still in the works?
Many in the storage industry anticipated Hitachi Data Systems would announce a refresh of the USP V architecture or add features to match those recently added by EMC Corp. in the EMC Symmetrix V-Max. "If you look at the size that V-Max can scale to now, EMC has pushed further," said Chris M. Evans, owner of Brookend Ltd., a U.K.-based storage and virtualization consultancy that works with EMC and HDS customers.
However, as both an EMC and Hitachi Data Systems customer, Thomson Reuters' Crowhurst said the EMC Symmetrix needed to increase the amount of available cache, and that's not a problem with Hitachi USP V. "The biggest thing about V-Max was the transformational step for the underlying architecture, which needed refreshing to continue to grow and scale," he said. "HDS doesn't need to refresh to overcome those limits yet." As far as further scalability and support for new features, Crowhurst said he sees both companies "on a natural leapfrogging cycle."
Hitachi Data Systems marches to its own drummer in the market, said Tom Trainer, founder and president at Analytico Inc. "My impression of HDS and where they're headed is that they will not react to V-Max in the way many would expect them to," he said. "I would imagine they have had plans to move beyond USP on the drawing board for some time on a certain schedule, and they're not necessarily going to modify that schedule in reaction to EMC."
Trainer said he would expect to see a bigger refresh from Hitachi Data Systems later in the year.
"I think they've got multiple things going on," said Rick Villars, vice president, storage systems and executive strategies at Framingham, Mass.-based IDC, of HDS. "To me adding this capability is something they needed to be doing and it's separate from any hardware refreshes. Customers have been asking for this functionality. I don't think it made sense for them to sit on it while waiting for a hardware piece."
The timing is right for another refresh soon, Villars said. "I expect to see more, but don't have anything definitive."
The Hitachi Data Systems USP V system launched in 2004, and five years is a long time without a new architecture. "We see our products as more evolutionary than something we have to refresh every so often," the firm's Mikkelsen said.
When it comes to the competitive picture between EMC and Hitachi Data Systems, Analytico's Trainer added that there's an elephant in the room—3PAR Inc.'s InServ Storage Server systems have included clustering and thin provisioning for years, while HDS and EMC have only recently added those capabilities. "I do see them following 3PAR's innovation lead," Trainer said.
New Sun array in keeping with cluster theme
Sun Microsystems Inc. also touched on the clustered storage theme with the rollout of a new model in its Sun Storage 7000 Unified Storage Systems line, otherwise known as Amber Road. The new Sun Storage 7310 scales from 12 TB to 96 TB of storage and can be configured in a dual high-availability (HA) cluster configuration. Previously, the only clustered model among the Amber Road products was the highest-end Sun Storage 7410, which scales up to half a petabyte and is priced starting at $57,490. Sun Storage 7310 pricing starts at $40,165.
Graham Lovell, Sun's senior director of open storage, said the Sun Storage 7310 was planned when the Amber Road products were first announced last November, but getting the higher-scaling model out first was a development priority. "We always planned to get back to cover the entry-level clustering opportunity," he said.
Sun is continuing with its storage hardware product roadmap while waiting for its merger with Oracle Corp. to close. It remains to be seen what Sun hardware Oracle will continue to sell. Lovell said he couldn't comment on any matters relating to the pending acquisition.