This brief is available in PDF; a synopsis follows.Download PDF

A snapshot is a point-in-time logical view of the data that is created by saving data differentials to a repository whenever the data in the source volume is about to be overwritten. The technique is also known as the “copy-on-write” technology that allows a snapshot to be instantaneously created while the source volume is actively serving I/Os. Essentially, the snapshot process creates an empty repository that will hold data differentials that are later changed in the source volume since the time a snapshot was taken. Snapshot images take up storage space only when the source volume receives updates that change the original data.