Snapshots • Snapshots are point-in-time views of data:
– Snapshot engine saves metadata that represents a view of the volume at a given point in time
– Metadata is used to build virtual representation of the volume as it existed at a point in time
– Several snapshots can be taken at different points in time – Uses minimal amounts of extra storage
Application server
Snapshot engine
+= Not the same size
Snapshot techniques are similar in concept to incremental backups: A data snapshot is simply a point-in-time view on a set of data, which can be created very quickly and without requiring large amounts of additional storage. In a virtual storage environment it works by copying the metadata describing the original storage, then writing updated blocks to another area of physical storage. Several snapshots can be taken at different points in time; the snapshot process needs to keep track of which physical data areas are used by which snapshots, so they can be freed for reallocation when snapshots are cleared.