Although there has been a sustained, high level of interest in the topic of Continuous Data Protection (CDP) in the storage industry for the last few of years, an authoritative definition of CDP remains a moving target. A survey of different vendors providing CDP solutions in the market today uncovers a wide spectrum of definitions. Some vendors intend CDP to mean the ability to restore data from any point in time. Others may describe it as the ability to get an application back on line quickly in a consistent form while minimizing the loss of data. Still others may define CDP as simply “snapshots plus data replication.”

A useful working definition is the one definition provided by the CDP Special Interest Group of the Storage Network Industry Association (SNIA). While the different notions of what CDP is and is not have prevented them to date from issuing a formal definition, they have been able to craft a working definition of CDP that is broad yet clear enough to be satisfactory:

“A methodology that continuously captures or tracks data modifications and stores changes independently of the primary data, enabling recovery points from any point in the past. CDP systems may be block, file, or applicationbased and can provide fine granularities of restorable objects to infinitely variable recovery points.”

Despite all of the differing ideas on the definition of CDP, there are a relatively clear set of conditions, characteristics, and metrics that set the boundaries for what a CDP solution must exhibit in order for it to be effective, or in other words, “true CDP.” First of all, it has to utilize RPO (Recovery Point Objective) as the key metric, by protecting data with a very fine level of granularity, since the CDP solution must be able to recover to any point in time. To do so, CDP solutions must be able to track data changes at the block, file, or application level, focusing down individual I/Os, data changes, or application events. As a secondary metric, the solution RTO (Recovery Time Objective) should also be low, meaning that the system can be restored to an uncorrupted point in time within a relatively short period of time.