InfoStor : Open source project challenges traditional storage, October 2006 Page 1

By Kevin Komiega

October 20, 2006—A number of open source storage projects have popped up in recent months, most with a focus on developing heterogeneous storage management software. But the latest open source group to make the scene is not aimed at building out the SMI-S specification or creating low-cost backup-and-recovery tools. The Cleversafe project is challenging the conventions of traditional data storage with an entirely new approach to how companies and individuals store, encrypt, and manage information.

Cleversafe's spokespeople make some pretty impressive claims. The technology under development enables the use of low-cost, commodity hardware to store data, capacity scaling is theoretically limitless, and data loss due to outages is a near impossibility. So how do they do it and why hasn't anyone thought of it before?

Cleversafe vice president of sales and marketing Greg Rudin says the rapid adoption of high-speed Internet connectivity makes the concept of dispersed storage a viable alternative to traditional storage architectures.

"There is an antiquated approach being applied to data storage today. The current methodology is based on storing multiple copies of the same data," says Greg Rudin, vice president of sales and marketing for Cleversafe. "You need ubiquitous adoption of high-speed connections to leverage the public Internet to move and store massive amounts of data. We are just now reaching that technology threshold."

The Cleversafe project is based on the use of information dispersal algorithms (IDA) and a grid system architecture to divide data into encrypted "slices" instead of making multiple copies of the same information.

In its current state the Cleversafe technology uses basic algebra to carve data into 11 encrypted pieces, each containing less than 10% of the original data. Each piece is then stored at a different remote data-center location. Rudin says this technique is inherently secure and reliable.

"This technology has been around since the 1970s and has been used to store things as sensitive as weapons launch codes. We're taking that methodology and utilizing it on a mass scale to provide uniquely cost-effective storage," he says. "The information is stored on 11 different servers in as many locations. We could lose up to five of those locations simultaneously and still recover the data."

He adds that the data is secure even in the event of a security breach. A hacker accessing a subset or slice of data at one location would not have visibility into the rest of the data slices at the other 10 remote sites.

Cleversafe launched its initial test grid last April with remote locations across North America in the US and Canada and plans to expand the project globally in the coming months. The ultimate goal is to create a storage Internet, much the same way TCP/IP paved the way for the public, open Internet that exists today and is supported by multiple ISP providers.

The commercial arm of the Cleversafe project hopes to begin offering software support and services to companies next year, but the basic IDA technology for building and operating a small storage grid is available for free under a General Public License (GPL) at Cleversafe.org.

The Cleversafe Open Source Storage Community was formed in 2006 to build and deploy new technologies for storing and securing data. The group's core of 20 full-time software developers is located within the campus of the Illinois Institute of Technology (IIT) in Chicago. Many of Cleversafe's contributors are students and former students of IIT.

Page 1 of 1