Riverbed plans to use some of its WAN optimization technology to free up space in existing data-storage systems and as a side effect speed up applications and servers that have to handle the data. (Learn more about WAN optimization.)

The appliance will be introduced during the first half of 2009 and will use Riverbed’s de-duplication technology to abbreviate data as it is stored and sent around corporate networks, Riverbed says. (Compare Application Acceleration and WAN Traffic Optimization products.)

Storing the abbreviations rather than the full data uses up less storage space, the company says. And because the abbreviated representations rather than the data itself are sent around the network in response to requests for data, fewer bytes get sent which reduces network traffic and congestion resulting in better performance, the company says.

For instance, if a law firm has 100 different contracts but all contain an identical indemnification clause, the Atlas will recognize the clause as a unit that is used over and over and index where to find it in the storage system, where it will be stored just once in its compacted form.

So the device saves space by storing data using fewer bytes than if it was fully written, and it also saves space by making it unnecessary to store the same byte pattern multiple times.

In reality, the device stores byte patterns as patterns, not in units that correspond to clauses, but the same de-duplication principle holds, Riverbed says.

The Atlas appliance has to work in conjunction with Riverbed’s existing Steelhead appliances that monitor traffic, break it down into byte sequences and create abbreviated references for those byte sequences. Atlas appliances index the references so they can call them out of storage to produce data that is called for by applications or servers.

Riverbed says Atlas appliances will only be deployed in redundant clusters to ensure constant data access. The Atlas gear is in alpha now and Riverbed says it is still figuring out what the appropriate deployment of Steelhead appliances is needed to support an Atlas.

The appliances can also be deployed in reclaim mode to free up space in existing databases. The appliance would sort through all the data, create references for byte sequences and re-store the data in its compacted form.