Patent attributes
A system and method for data backup. A computer system includes clients coupled to backup server and single instance store via a network. Prior to requesting an initial data backup, a new client segments and fingerprints the data to be backed up. The new client creates a sorted list of hash values and removes duplicate values in the list. In response to receiving the sorted list, a backup server may further reduce the sorted list to a unique list by removing hash values corresponding to data segments already residing on the single instance store. In response to receiving the unique list, the new client may transmit corresponding data segments to the backup server via the network. The transmitted data segments are added to a single instance store.