Users drive open source storage initiative

By Kevin Komiega

Another storage industry group has joined the race to deliver a universal storage management framework, but this time the vendor community isn’t leading the effort. Last month, a Website called StorageRevolution.com went online with a mission to solicit ideas, resources, and code from the end-user community in an open source effort to advance heterogeneous storage management. And if the mission statement and opinions available on the StorageRevolution.com Website are any indication, the group could cause some waves in the storage pool.

StorageRevolution.com is seeking to build a framework to connect application-aware storage management tools and already has a template in place to facilitate the process. The proposed management framework, called CAPSAIL, is aimed at helping storage managers understand the requirements of their applications and to automate the provisioning of storage resources and protection services to application data.

The goal for StorageRevolution.com is for CAPSAIL to ultimately enable users to “plug in” a mix of best-of-breed commercial and open source tools for more-efficient storage management. Users of storage technology can participate in the development of CAPSAIL for free by sharing their management requirements via the Website. Users with design and coding expertise are also being asked to participate in modular development projects. The resulting CAPSAIL framework will be available free of charge under a standard GNU/GPL Open Source License, according to StorageRevolution.com.

Toigo Partners International
Click here to enlarge image

Jon William Toigo, founder of the group and CEO of Toigo Partners International, says vendors are invited to participate in the StorageRevolution.com project, but their involvement is not required for the open source initiative to be successful. The user group is already working on the CAPSAIL technology.

“This is a standard open source initiative,” says Toigo. “What we have on the table right now is a technology called Spider Web. It’s a token-based messaging scheme for communicating between nodes in a peer-to-peer network and was developed by a group of storage industry veterans.”

Toigo says Spider Web is the foundation of the proposed CAPSAIL management framework and is applicable for collecting information about disk arrays and how data is being generated by applications and storing it in a common repository.

Toigo admits that Spider Web is not nearly as developed as the Storage Networking Industry Association’s (SNIA) Storage Management Initiative Specification (SMI-S), but he contends that the technology developed by StorageRevolution.com members may be the only option for small software vendors.

“SMI-S was set up to freeze the little [ISVs] out and only include vendors with the resources to develop providers,” says Toigo. “Users are not demanding SMI-S technology. It is not a checklist item [that customers require].”

The SNIA’s SMI-S is a storage management standard based on the Web-Based Enterprise Management (WBEM) architecture and the Common Information Model (CIM). SMI-S provides a common interoperable and extensible management transport, an automated discovery system, and a unified object model for the control of storage LUNs and zones in SANs. Native SMI-S providers are CIM-based device management interfaces that allow storage devices to connect to, and be managed by, any SMI-S-compliant management application.

Toigo argues that storage industry associations and their vendor membership have little to gain by backing a cross-platform management solution.

“StorageRevolution.com has no disagreement with SNIA’s fundamentals and view. We disagree with the approach to the problem,” says Toigo. “I feel that the SMI-S effort has hit the skids and is mired down in industry politics because the industry is trying to accomplish the development of a unified storage management framework through ‘co-opetition.’ ”

SNIA’s chairman and EMC senior technologist Wayne Adams does not agree with the statement that the aim of SMI-S is to “freeze out” smaller software vendors, nor does he believe that StorageRevolution.com is intent on replacing SMI-S as a standard interface. Rather, Adams is optimistic that the group will make use of the SMI-S interface and other existing standards and development tools from SNIA.

Adams says, “Overall, the way SNIA looks at its role in the adoption of storage networking technologies is ‘the more, the merrier.’ We believe an offering like what [StorageRevolution.com is planning] could help provide the next generation of code base. If it’s robust enough for the task it should be phased in. But regardless of how the market evolves you’re still going to need specifications and interfaces because there is so much equipment out there. SNIA’s role is to provide those interfaces.”

Adams believes that the advent of open source initiatives for the storage industry will complement SNIA’s ongoing development of management tools and standards for interoperability between storage applications and devices. However, Adams also believes that the development of a storage management framework is a sizeable task and requires users and vendors to rally around open source technology for it to take hold.

“If you look at past acquisition and merger activity you’ll see that if a bunch of smaller ISVs come together and develop a framework it might take a larger vendor to provide end-to-end solution sets [and support],” says Adams. He likens the process to the evolution of the Linux operating system and the manner in which companies like SuSE Linux and Red Hat now provide services, support, and product suites based on freely available open source Linux technology.

Adams also says that any management applications or middleware developed by StorageRevolution.com contributors will require extensive testing in heterogeneous storage environments, which is another way that SNIA could contribute to the open source cause.

“If the StorageRevolution.com project moves forward and they are successful and deliver something to the market there will still be some level of necessary application testing,” says Adams. “The SNIA technical center facilities offer vendor-neutral test beds that are broader than what any one company can provide.”

Toigo addressed the interoperability testing issue by committing to making the test labs at Toigo Partners International available to StorageRevolution.com developers at no cost.

StorageRevolution.com is the latest open source storage group to materialize, but it’s not the first. An IBM-led group of nine storage hardware and software vendors joined forces in October 2005 to launch Aperi, a non-profit, open source community aimed at developing a heterogeneous, standards-based, common storage management platform. The difference? Aperi is based on vendor contributions while StorageRevolution.com is aimed at recruiting end users.

The Aperi group issued a statement in response to the launch of Storage-Revolution.com in which the group also shared SNIA’s optimism for a symbiotic relationship among all of the storage industry associations.

The opinion of Aperi and its membership, according to the statement, is that the open source community, as well as similar industry initiatives coming out of StorageRevolution.com and SNIA, have set out with “the common goal of providing customers with greater flexibility in the way they manage their storage environments.”

The statement also says Aperi is still in the process of fine-tuning how its organization will operate and what it will deliver.

The group has told the media that the Aperi platform will build on SNIA’s SMI-S standard. SMI-S will serve as the common interface between heterogeneous storage devices, while code developed through the Aperi project will provide a framework for applications to perform high-level management functions.

Aperi’s initial list of members included Brocade, CA, Cisco, Engenio, Fujitsu, IBM, McData, Network Appliance, and Sun. Conspicuously absent were vendors such as EMC, Hewlett-Packard, and Symantec, which have not yet joined the Aperi group.

This article was originally published on March 01, 2006