I know in the past snapshotting was used for availability rather than backup. What is driving interest in using the technology as part of a data backup strategy?
The fun part about snapshots, other than the near instantaneous rollback of data and some other data management agility capabilities, is that the debate of backup technologies between “snapshots versus backups” is finally over. I wrote about that in my blog earlier this year, but the short version is that the storage folks used to advocate snapshots because they didn’t affect server performance and they roll back so quickly. But they historically haven’t always done as well for transactional applications — or granular file recoveries. Alternatively, backup software does really well with transactional apps and granular recoveries, but restores can take longer than rolling back from a snapshot and can incur I/O penalties.
Today, we are seeing a lot more examples where backup software is being integrated with array management tools, with the result being a truly better together solution. For example, Symantec NetBackup has a feature which enables it to manage the schedules of when snapshots occur, and the snap catalog is transparently integrated into the storage pool. The result is that you can choose a file to restore, and where before it might have come from disk or cloud or tape, now it could come from a snapshot and, as the administrator, you can’t tell the difference and don’t care.