Monday, December 22, 2008

The 'C' in CDP has been downgraded

I’ve been doing some background research on continuous data protection (CDP) for an upcoming Special Report, and I’m starting to wonder whether the term – not the technology – will survive.

For one, we still have the “true CDP” (recovery to any point in time) vs. “near CDP” (snapshots on steroids) argument confusing things. I spoke to a number of vendors, and they all went to great lengths to explain why the distinction is irrelevant, and then they went on to explain why their approach was (or in some cases, wasn’t) true CDP.

Maybe it is meaningless. The fact is that very few companies really need true CDP. It just depends on how much granularity you need in your recovery point objective (RPO), and for many companies snapshots are more than adequate.

In addition, many vendors now offer a spectrum of recovery points, ranging from snapshots to APIT recovery: Users can choose based on their specific RPO requirements, capacity issues, cost, service level agreements (SLAs), “application-consistent” vs. “crash-consistent” requirements, the importance of the data being protected, etc.

Further, many traditional backup software vendors have integrated some level of CDP functionality into their backup suites, or at least offer it as an option. This blurs the lines between CDP and backup and, in fact, standalone CDP applications/platforms are becoming rare.

As such, it’s possible that the CDP term will eventually fade away, despite the fact that the Storage Networking Industry Association’s Data Management Forum has a special interest group dedicated to CDP.

As Fadi Albatal, FalconStor Software’s director of marketing, says, “The argument [over true CDP vs. near CDP] is irrelevant. It all depends on your SLA, RPO and RTO requirements. It’s a matter of what is acceptable risk and/or data loss.”

Or as John Ferraro, InMage Systems’ president and CEO says, “We don’t sell CDP; we focus on disaster recovery.”

No comments: