Feature

Cleversafe envisions a new world of storage

Imagine a world in which there are no storage vendors -- at least not vendors as we know them today. Imagine a world in which storage, like bandwidth, water or electricity, is a utilitarian resource; a world in which storage and the Internet work the same way and can even be one and the same thing.

No, it's not a song John Lennon wrote -- it's the vision of Chris Gladwin, CEO of Cleversafe, an organization working on developing an information-dispersal algorithm (IDA) that he claims could change how we think of both storage and computing.

The Cleversafe concept has been around since the 1970s in one form or another, Gladwin said, but the model is only becoming practical to implement now that high-speed Internet has become ubiquitous.

IDA makes "slices" of data that contain added code that "scrambles" the data for encryption, as well as corresponding to a kind of parity slice called the Coded Slice, which allows the data to be rebuilt even if some nodes are lost. In other words, according to Gladwin, it's like geographically dispersed RAID, with encryption built in, that works over a commodity Internet connection.

Gladwin said it's the parity that makes it different from standard grid computing. "Any majority of nodes can put the data back together -- in the case of an 11-node system, six nodes can recover all the data because of the algorithm," he said. "Both conventional RAID systems and conventional grid systems can't withstand losing five of 11 nodes or disks."

In fact, he said, if an 11-node grid was built using commodity servers with 99.9% availability, it would have 12 "9s" of reliability for the entire system -- "an outage of one hour in a million years," Gladwin said. "I thought that had a certain ring to it."

Right now, the Cleversafe Research Storage Grid consists of 11 nodes -- the maximum number of nodes that the current IDA allows. Theoretically, the algorithm can be expanded infinitely, Gladwin said, but right now that expansion is the challenge Cleversafe was founded to address. "We're working on prototypes of 30-, 40- and 50-node systems," he said.

Imagine all the people, sharing all the world

Currently, the organization working to advance the concept has two sides: Cleversafe.org, an open source project that's developing the code, as well as the bigger prototype systems and Cleversafe LLC, which aims to commercialize the concept once it's ready for prime time.

Ultimately, according to Gladwin, the vision of Cleversafe's 20 open source developers is a worldwide storage grid, which would be run and controlled by Internet service providers (ISP). "A megabit of bandwidth and a terabyte of storage could be the same thing," Gladwin said.

It may seem far-fetched, and Gladwin admits the storage Internet won't exactly be coming next year. "The limiting factor is time," he said. "There's other behind-the-scenes stuff, like LDAP (lightweight directory access protocol) for the IDA so it could work over the conventional Internet, that have to be developed, and the self-healing capabilities that could come with the algorithm are also being worked on now. "

But Gladwin, it turns out, has already had a hand in several computing trends that seemed harebrained at first and have since become ubiquitous. The most recent item on his resume prior to this project is MusicNow, which provided digital music services marketed by Best Buy, Clear Channel Communications, Microsoft, SBC Internet Services, Charter Communications and EarthLink Inc. MusicNow grew to 100,000 customers before being bought by Circuit City Stores Inc. for an undisclosed sum in early 2004. Circuit City then sold MusicNow to America Online LLC (AOL) in November 2005. AOL currently uses MusicNow as its music-sharing service for subscribers.

Before MusicNow, Gladwin was the primary inventor on the original patent for the wireless thin client. He served as the founding chairman, president and CEO of Cruise Technologies, which provided wireless thin clients that were marketed by distribution partners, such as Motorola Inc., IBM, Wyse Technology Inc., Telos Corp. and Zenith Data Systems.

Could it be he's on to yet another something?

He could be, according to Arun Taneja, founder and analyst with the Taneja Group. Though Taneja said he's seen companies with similar ideas fail because they could not get funding for such radical concepts, the previous prototypes he's seen came along several years ago, and times may have changed in Cleversafe's favor, he said.

"What's interesting here is the connection with the Internet. I think they could build up a customer base from consumers and small enterprises first that way," he said.

In terms of the enterprise, Taneja said, there are a few factors in Cleversafe's favor over previous distributed-computing prototypes. One is that the algorithm allows for security, by encrypting the slices of data in flight and providing for a "key" system so that only the owner of the data could put the slices back together again. Another is that applying it to the Internet would take advantage of existing storage and networking resources, not totally unlike previous prototypes, but on a much broader scale, Taneja said,.

As such, this kind of revolutionary concept might not take over in the next year or even the next five years, but sometime within the next decade. It could very well become the next storage standard, Taneja said.

"It may take some time to get people to trust the idea of breaking up the data into these small chunks," Taneja said. "But in reality, it's more secure than having data on tapes falling off of trucks, and, if they can truly prove the concept, would probably improve storage utilization rates as well."


Email Alerts

Register now to receive ComputerWeekly.com IT-related news, guides and more, delivered to your inbox.
By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

This was first published in December 2006

 

COMMENTS powered by Disqus  //  Commenting policy