Five months after acquiring the intellectual property assets of MAID maker Copan Systems Inc., Silicon Graphics...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Inc. (SGI) is re-releasing the technology under its brand as the Copan 400 and claims it has boosted the performance of the disk spin-down system.
SGI claims the new Copan 400 can achieve 6,400 MBps throughput through a native disk interface thanks to the addition of 8 Gbps Fibre Channel (FC) links and Intel COM Express quad-core processors that run at approximately 3.0 GHz. The Copan virtual tape library (VTL) model, based on a partnership with FalconStor Software Inc., will reach 3,200 MBps, according to SGI data protection product manager Jim Miller, because SGI needs to work with FalconStor on getting the VTL integration with Copan to support 8 Gbps Fibre Channel.
Customers will also have to wait for the new performance rate through native disk interfaces because SGI is still working on integrating disk-to-disk APIs with IBM Tivoli Storage Manager (TSM) and CommVault Systems Inc. Simpana backup software applications. Miller estimated those integrations will be done in approximately two months, with more backup software partners to follow.
SGI also plans to expand the Copan 400's host interface options, beginning with iSCSI sometime before the end of the year and moving on to InfiniBand and others through modular interface cards that don't require replacing the whole system.
Miller said when it comes to performance, "the shelf itself within the system was the limiting factor," and that has been updated with "off-the-shelf components" that perform faster. SGI is keeping Copan's non-standard disk array shelf, designed to pack hundreds of drives into a single rack, for now because of the density it provides.
Wikibon.org analyst David Vellante wrote in an email to SearchStorage.com that in his view, the main reasons Copan wound up on the auction block in the first place earlier this year were performance issues, and the fact that integration was required for applications to use the spin-down array without timeouts (this was also the consensus on Copan at the time of SGI's acquisition in February). "So increase the option to allow for 50% of the [disks] to be active and you still have the same problem, just at a smaller scale," Vellante wrote. "Maybe in SGI's markets this plays just fine because of sensitivity to [energy] sustainability. but I don't think this changes the game by any means [other than as] another option for SGI customers."
Customer says support contract issue a deal-breaker
These changes will come too late for one customer, who said he's planning to swap out Copan with disk arrays from Nexsan because SGI will not assume support contracts for those who bought the product prior to SGI's purchase of Copan's IP.
Dean Flanders, head of informatics at the Friedrich Miescher Institute (FMI) in Basel, Switzerland, said at the time of the SGI-Copan acquisition the non-standard shelf (which made expansion costlier) and performance of Copan's MAID product already had him considering alternatives. But this week, Flanders said he made his final decision after SGI told him it wouldn't honor his Copan support contract.
"I paid Copan for four years of service," he said. "SGI won't honor that contract — I have to repurchase support from SGI." Saying he has "absolutely no allegiance to keep that gear," Flanders plans to go with Nexsan's disk arrays with AutoMAID instead. Those arrays also use non-standard shelves, he said, but they fit in a standard rack, and can offer between 400 TB and 1 PB in a rack, enough density to meet his needs. "I still think Copan is a great idea, but between all the issues and risks, I don't think it makes sense to stick with it," Flanders said.
SGI responded through an emailed statement from a spokesperson: "When SGI acquired Copan, it acquired only assets and not liabilities. Liabilities would include support contracts. SGI is actively working to resolve support issues on a case-by case-basis to minimize the impact on customers."