eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Last year, while extolling the promise of its enterprise storage server—code-named Shark—IBM officials said they would later add to it virtualization capabilities.
However, late last month, IBM quietly nixed plans to implement virtualization in Shark. Instead, the company plans to develop an enterprisewide virtualization technology that sits on the network close to the switch—in an area also known as the “SAN [storage area network] cloud.”
IBMs change of direction underscores a growing debate in the storage industry revolving around virtualization and the frustration of IT managers anxiously awaiting the technologys promise of interoperability between storage systems and networks.
With storage virtualization, IT managers would have the ability to logically move terabytes of data with a few simple keystrokes as well as more easily manage the vast and disparate storage devices they oversee.
But theres disagreement among vendors as to just where within the storage framework the virtualization should reside—on the server, within the SAN cloud or in the applications. As a result, IT managers are worried that storage vendors from separate camps will end up squaring off.
The result could be the same problem that plagues other areas of storage: technology from one vendor that does not communicate with another vendors products.
“I hope these companies are talking to SNIA [Storage Networking Industry Association] and the standards bodies as they are developing these products,” said Rick Bauer, CIO of The Hill School, based in Pottstown, Pa. “I have five different network-attached storage products, and none of them talk to each other, and none can see each other.
“There is no single tool that virtualizes them all,” Bauer said. “That is the virtualization Holy Grail Im looking for. Everybody wants virtualization, but then there are the NIMBYs, those who dont want to provide the hooks into other storage.”
Compaq Computer Corp. is working on a virtualization software piece called VersaStor, and both it and IBMs Tank are expected out by years end. Earlier this month, Kom Networks Inc., of Nashua, N.H., announced its Virtual Storage Works, which virtualizes storage at the file level.
Storage service providers such as NaviSite Inc. want to see virtualization done at the application level, so the application decides to move storage based on where it is needed.
“I want to solve the problem on the software layer, by making the application aware at the file level of the storage assets,” said Peter Kirwan, vice president of business strategies at NaviSite, in Andover, Mass.
When Shark was first delivered about 18 months ago, IBM, of Armonk, N.Y., planned to put virtualization within the disk array. Now the focus is on Tank, a policy-based management virtualization for data sharing on heterogeneous server platforms. Internally, IBM did not find a compelling reason to put virtualization in Shark, according to analysts. For example, the company found a different way to execute the snapshot copy feature.
“Those [customers] that were expecting [virtualization with Shark] Im sure are disappointed. But I actually dont think many people were expecting it because IBM was soft-pedaling it from the start,” said John Webster, an analyst with Illuminata Inc., in Nashua.
IBM declined to comment, instead referring to its press release issued late last year outlining plans for Tank.
