Deduplication vs. Compression at Backup
Both the deduplication and the database compression aggregates data. In case of a compression, a determined amount of data are searched for frequently occuring patterns which may be expressed in a better way. For example, a binary pattern with 8 subsequent zeros or ones might be substituted by 3 zeros or ones. The average compression rate of a dataflow is limited by a classical backup to 2:1, 3:1 at most. With the deduplication, the current full backup is compared to the previous one. Due to the relatively small change rate of 2 up to 10%, the data amounts are quite similar. With the increasing number of data backups, the similarity of the on block level increases even more so that compression rates of a factor 9 to 20 can be achieved.