Data backup speed and what are the factors

In a recent interview, an insurance company user pointed out that the backup of a large-capacity file system composed of small files, if the backup speed can reach 80gb/hours, is quite good. To this end, I looked at the data, taking the HP tape library as an example, the maximum can be reached 38tb/hour level. The insurance company is also using LTO3 tape drive Library, why is there such a big gap in the speed of backup?

To this end, I consulted the industry-related experts. Experts point out that the speed of data backup does not depend entirely on the speed of the tape library or tape drive, but also on the backup method of the backup software. The backup software, he points out, is backing up the data the first action is to scan the data that needs to be backed up, taking the insured data as an example, with millions of small files, and if the backup software needs to know about the small files on the disk, and then copy to the backup server over the network, This is a fairly time-consuming operation, where the performance of file servers, networks, and backup servers can be a bottleneck for speed. In this case, some backup software uses a "volume copy" to copy the entire volume that needs to be backed up, such as copying a 20GB C disk, regardless of how the small files are stored on the C disk and copied to the backup server. All of which are "volume copies" and there are subtle differences. Some backup software "volume copies", regardless of how much space the actual data occupies, copies the entire volume. But good backup software, such as CommVault Simpana, is just a "volume copy" of the actual data.

The process of "volume copy" is the process of reading data. After the data is read to the backup server, the backup software copies it to the tape library or tape drive. Experts point out that the recovery speed of backup data is critical, after all, the purpose of the backup is not a simple backup, the speed of the backup is to restore, if the recovery is too slow, then the efficiency is greatly compromised. He pointed out that the speed of recovery, the core of the backup software to tape copy of the way the data. In the case of a volume copy, if the backup software is simply a swallowed copy of "volume copy" data to tape, then the recovery must be all data. Excellent backup software, can do a single small file, that is, fine granular file recovery, it does not recover all the data, can do to directly restore the target file. As a result, different backup software has a very large impact on the speed of backup, and more crucially, there is a very large difference in the speed of recovery, which users should carefully analyze and compare.