Efficient Metadata Management in Net-Oriented Block-Level CDP System.

Continuous Data Protection (CDP) is an emerging scheme of data backup and recovery. Compared with traditional data protection techniques, it can provide much higher reliability and allow the state of the storage to be reverted to any point-in-time (PiT) in the past. One big challenge which affects CDP recovery performance most is how the metedata of CDP are managed. During our experience of CDP system developing, we observed that efficient metadata management is so critical to the overall performance of CDP system. To illustrate how different management methods contribute to the CDP system, we present three related methods in this paper and evaluate all of them empirically. Our experimental results demonstrate that the difference in performance may amazingly reach several orders of magnitude. Our main contribution is twofold: we have put forward to two recovery algorithms called Full Recovery and Incre Recovery and three implementations called DIR_MySQL, OPT_MySQL and LOG_CDP for each, in particular, the one that is much faster than the other two. [ABSTRACT FROM AUTHOR]