Storage snapshots offer capabilities beyond those of typical backup software. Whereas a full-volume restoration might take the better part of an hour to complete, a snapshot can be rolled back in a matter of seconds. Furthermore, some of the next-generation snapshot products provide rollback capabilities for individual files and applications. In this video, independent backup expert Brien Posey discusses how snapshots work and where the technology might fit in your data protection strategy.
Snapshots aren't a copy of your data. Instead, they act as representations of data from specific points in time and can roll back an application to that point. While they can't replace a backup directly -- snapshots without data don't work and, frequently, snapshots are stored on the same volume as their associated data -- they can be used to augment traditional backups.
There are two main types of storage snapshots in use today -- differencing disk snapshots and pointer-based snapshots. Differencing disk snapshots are based on the use of differencing disks. The accumulation of these snapshots tends to degrade the server's read performance.
Pointer-based snapshots are based on the use of pointers to storage blocks instead of differencing disks. They do not incur the performance hit that differencing disk snapshots do, but they result in far more storage consumption.
Potentially, snapshots can serve as a backup option, but administrators need to plan ahead. First, administrators have to be sure that they conduct block-by-block mirroring or replication of both the snapshots and the associated data.
Regardless of which type of storage snapshot an organization decides to create, it is critically important to make sure that the snapshot software supports any applications for which you intend to create a snapshot. If it does not, data corruption can occur.