A method for compressing data by storing only changes to the original data. For example, when a document is created, it is stored as a regular, complete file. When changes are made, deduplication stores only the data that has been modified in a subsequent file with indexes to the original. The changes may be identified by bytes or blocks.

In file deduplication, only one copy of the file is stored even though it may be referenced from many sources. For example, if the same attachment is used in several e-mail messages, the second and subsequent messages are changed to link to the original attachment rather than duplicating it each time.