Can anyone please explain the concept of snapshot in emc?

Snapshotsallows companies to make more effective use of their most valuable IT resources by enabling parallel information access. Instead of traditional sequential information access—forcing applications to queue for information access—Snapshots allows multiple business processes to have concurrent, parallel access to information.

Two different views of your volume can be created. First, snapshots are block-based logical point-in-time views of production information. Using only a fraction of the original file system space, multiple snapshots can be created to suit the need of multiple business processes. Secondary servers see the snapshot as an additional mountable disk volume (or LUN). Servers mounting a snapshot have full read/write capabilities on that data.

It also has the ability to create a block-based, full mirror copy of the source volume, called a Clone. This full-volume copy is identical to the original, and can be written to, just like a regular volume. After an established period of time, it is possible to incrementally re-establish or restore the clones with its source volume.

A snapshot is not a copy of your information; it is a logical view of the original information based on the time the snapshot was created. A clone, on the other hand, is a full-image copy of your data. Snapshots and clones can be created in seconds and can be retired when no longer needed.

Thanks Shijas..That was very helpful..

Excellent answer, and absolutely true if you are only dealing with Open Systems environments. The one place you need to be careful (which I learned the hard way) is with a mixed environment of Open Systems and Mainframe on Symmetrix. Apparently there is a function for mainframe environments called a dataset snapshot (using TimeFinder) which is actually what we (in the Open Systems world) would call a clone. It's a full physical copy.

Most people won't have to worry about this, but I bring it un in case you find yourself in an argument with the mainframe storage folks. Due to some poor choices/consistency in feature naming, they may actually be right this time around :-)