Data protection can occur locally and remotely. Backups allow for both physical and logical data protection. The first location of a backup is traditionally at a local site. However, having a second copy of the backup at a remote (that is, at a far enough distance to avoid a regional disaster) disaster recovery (DR) site is a good idea. Manually moving tape copies off site is a tried-and-true method, but the length of time before applications can be restored (which may be days) is increasingly unacceptable in an always-on world. Many backups are now done to disk rather than tape using a disk-based backup approach such as virtual tape library (VTL).
Another alternative is to use electronic remote replication as opposed to manually transporting tapes. The problem with using a network to back up over a distance has historically been not only the bandwidth cost but also the time that takes to do the backup (in addition to the cost of the array on which the backup is stored). Companies can reduce the amount of data that has to be transmitted each day (and hence the time it takes) through the use of data deduplication in conjunction with a VTL. That would indicate--in theory, anyway--that electronic remote replication can be cost effective. (Note that, for this discussion, replication does not imply mirroring, which allows applications to be restarted from a remote array. In this discussion, replicated data still has to be restored before it can be used.)
Survey Sheds Light on Remote Replication
Sepaton--which provides data protection and VTL products--conducted a survey to examine the remote replication and disaster recovery priorities for large enterprise backup environments. This is the fourth annual survey that Sepaton has commissioned for North America and Europe across a number of verticals. The size of the backups was notable: more than half (55%) have full backup volumes of more than 50 Tbytes, and 14% have full backups of more than one petabyte.
Next: Active and Passive Approaches