Previously identified by Network World as "One of 10 Open Source Companies to Watch," Cleversafe has just released a commercial version of its Dispersed Storage Network (dsNet). In 2007, Network World called dsNet one of six advanced-technology open source projects likely to become the "Best of the New Data Center." With glowing reviews like that, how can you not be curious about what Cleversafe now has to offer?

Cleversafe has taken the principle of the Internet and applied it to storage; that is, it’s like a packet-switched network for data storage. Rather than simply making a copy of data and putting it on another server somewhere, the Cleversafe approach uses “information dispersal.” Data is run through an algorithm that slices the information to create representations of the original data. Those slices are then distributed across a grid network, the dsNet.

The slices contain additional information that allows you to recreate the original data 100%, even if some of the slices are unavailable, say from a hardware failure. If a slice goes missing or becomes corrupt, it can be automatically rebuilt using the other slices. This rebuilding process goes on in the background, checking on the status of all the slices to continually ensure high availability.

The open source version of this product has been available and in use for about three years. The technology is proven and ready for commercial use. Cleversafe says the ideal application is “digital content loading,” meaning digital videos, photos, other media and unstructured content.

Here’s a quick look at the three main components of a dsNet.

1. CS Slicestor (Dispersed Storage server) - This is the device where the data slices are stored. A single server offers 3 terabytes of “raw” storage per 1U slice server. In a typical configuration, you can network Slicestors together to get up to 30 terabytes of usable storage. There is approximately a 30% to 60% overhead, depending on the threshold you set for data rebuilding. Slicestors can be deployed in a single rack or geographically distributed around the world.

2. CS Accesser (Dispersed Storage router) - This component slices, disperses and retrieves data to and from the Slicestors. It holds the “secret sauce” – the algorithm that converts raw data from your applications to the representative packets called slices. It has iSCSI or Block interfaces, and can be deployed in redundant configurations.

Linda Musthaler is a principal analyst with Essential Solutions Corporation.