Making a case for CAS

ONCE AGAIN EMC has spotted a good idea and made the idea its own. In the same way that EMC took a hint from crafty storage software startups last year and then all of a sudden decided to become "the first" true enterprise storage software vendor with its AutoIS storage management initiative, EMC has formed a new technology plan on the shoulders of NAS startups and metadata technology developers called CAS (content addressed storage).

Free IT resource Virtualization Insights from Top Experts - Learn how virtualization gets real! Sponsored by Dell Free IT resource TechNet: More ways to know it, share it, and keep it running. Sponsored by Microsoft

That's right, just when we were all waiting for another confusing storage technology acronym to come along, EMC has leveraged its acquisition of FilePool to bring us CAS, a storage platform that speeds the delivery of large files stored in colossal databases.

The concept behind CAS is to provide an intelligent repository able to store and retrieve information independently from the application or the operating system. That repository, which arrived in the form of EMC's new Centera CAS server, assigns to each new data object a fixed location in the repository and a unique identifier that client applications use to retrieve the information. The resulting content address is similar to a digital fingerprint for each file, which makes it easier for the Centera storage server to retrieve them.

This method of storage is a drastic departure and a much-needed simplification from SAN and NAS environments, where the applications decide where information is stored. In the CAS world, applications use a simple programming interface to communicate with the storage system, leaving storage allocation and retrieval issues to the storage system.

Because EMC's Centera server essentially hangs off the network and serves files, CAS technology could have arrived as the "killer app" for NAS systems, or a gateway architecture for the convergence of SANs and IP networks. But now we must find a place for our CAS next to our SANs and NAS devices.

It's reasonable to argue that different kinds of data -- electronic purchase order, spreadsheet document, digital photo, etc. -- call for different storage strategies, and CAS addresses the problem of simplified management of files with fixed content, such as digital images and PowerPoint presentations. But do we really need yet another class of storage when the industry's momentum is toward consolidation of storage management tasks? EMC seems to believe that we do.

In fact, EMC also believes that CAS deployments, such as AutoIS deployments, are best optimized through the commitment of third-party APIs to the platform. With CAS and Centera, application vendors are required to write or re-write their applications to include Centera-compliant APIs if they'd like their software to take advantage of Centera's file access speed. As selfish as that may seem, EMC will likely have little problem pipering APIs to its doorstep, because few apps vendors will want to be left out.