This article can also be found in the Premium Editorial Download "Storage magazine: Upgrade path bumpy for major backup app."
Download it now to read this article plus other related content.
Many experts say CDP will replace traditional backup. But before you take the plunge, here are some key points to consider.
Continuous data protection (CDP) is getting almost as much press as Apple's iPod. CDP tracks data modifications and stores changes independent of the primary data, and lets you recover data in seconds from any point in the past. Some pundits have postulated that CDP will replace backup. But is that just new technology hyperbole or is CDP a must-have technology?
CDP provides a recovery point objective (RPO) and recovery time objective (RTO) of essentially zero data loss with a very fast recovery. Defined another way, CDP is a time-stamped backup stored on secondary disk. The appeal of CDP is its ability to quickly rewind applications to any point in time to find a consistent image of the data.
But don't confuse CDP with mirroring, which provides data protection only from hardware failures. If data is corrupted or deleted on the primary system, it will be on the mirrored copy, too. CDP provides protection from both hardware and data failures.
CDP is also sometimes confused with so-called "fine-grain" snapshots. Snapshot products capture changes as points in time, with every snapshot checked for consistency before the next one is taken. This is analogous to a digital camera rapidly taking multiple photographs vs. a video camera capturing every nuance of the entire sequence. There are time gaps between snapshots,
CDP ensures there aren't any gaps because it captures every file, block or table change as it occurs. And while it's possible to restore data to any point in time with CDP, there's a consistency issue because it's impossible to determine which point in time to rewind to. Until recently, most CDP products weren't able to identify the most recent point in time when the data was verifiably consistent.
So restoring CDP data has become a trial-and-error process; an administrator must guess a point in time to restore from. If the guess ends up being a recovery point after the corruption occurred, the data must be recovered again from an earlier point in time--greatly increasing the recovery time. If the administrator plays it safe and chooses a recovery point too far back before the known corruption, the CDP recovery time can be worse than when using frequent snapshots or backup to disk. This guessing game to find the exact point in time from which to recover can nullify the fast recovery CDP is supposed to provide.
This was first published in September 2006