A method and device for realizing incremental deduplication among multiple snapshots of remote replication

Abstract

The invention discloses a method and a device for remote copy. The method comprises the steps as follows: snapshots of a plurality of adjacent time points are generated, a change of content between the snapshots of any two adjacent time points is an increment, increment de-weighting is performed on the two adjacent increments, and increment results after the increment de-weighting are copied remotely, wherein the increment de-weighting comprises at least one of the following content: if the two adjacent increments are the changes of data, the same data are merged; if the two adjacent increments are the changes of metadata, the increments established and deleted for the same file or catalogue are offset, or the increments modified for the same file or catalogue are merged; if the two adjacent increments include the increments deleted for the file or the catalogue, the other increments for the file or the catalogue are deleted; and if the two adjacent increments include the increments established for the file or the catalogue, the other increments for the file or the catalogue are deleted. According to the method, repeated or unnecessary copy operations can be reduced, so that the backup efficiency is greatly improved, and bandwidth resources are saved.

Description

technical field [0001] The invention relates to the technical field of virtual data protection, in particular to a method and device for realizing incremental deduplication among multiple snapshots of remote replication. Background technique [0002] Remote replication is a commonly used data protection method in the storage field. Through remote replication operations on the production end and disaster recovery end, multiple copies of the same data can be generated for one piece of data. Snapshot is a very important method for system disaster recovery backup. The main function of the snapshot is to realize the online backup and recovery of data, which can quickly restore the data to the state at a specified point in time when the system fails, and allows users to access the snapshot data at any point in time in a read-only manner. [0003] The mainstream methods to implement snapshots include COW (Copy On Write, copy on write) and ROW (Redirect OnWrite, redirection on writ...

AI Technical Summary

Problems solved by technology

In fact, node d is created at snapshot N and deleted after N+1, so for the disaster recovery side, d between snapshots [N, N+2] is equivalent to never existing, and the newly created increment and deletion increments can offset each other, the creation and deletion of node d on the disaster recovery side is an unnecessary copy operation, wasting bandwidth resources

[0005] Therefore, in the file system, traditional remote replication performs data synchronization through replay between snapshots, resulting in more repeated operations, reducing system performance and backup efficiency, and wasting bandwidth resources.

Images