Ca, Inc.

Liqiu Song

Chandra Reddy

Prashant Parikh

Ming Yan

A computer implemented method and system obtains current signatures of data chunks and performs a proximity search of a library of previous signatures as a function of the likely location of corresponding data chunks. A full search of the library of previous signatures for those current signatures not found in the proximity search is also performed.

Proximity and in-memory map based signature searching for duplicate data

The various embodiments herein include operate to identify, consolidate, and reduce redundant backup data storage. One embodiment includes storing data blocks and first signatures of data chunks of each stored data block, the first signature of each data chunk including a reference to a storage location of the data chunk within a stored data block, the stored data blocks including data blocks of previous and recent backup sessions. Some embodiments further include storing second signatures in a second signature repository, where the second signatures are calculated based on determined boundaries of the first signatures from previous backup sessions. At least one of the second signatures is calculated based on at least two first signatures, and in the range of 32 to 64 first signatures in some embodiments. Some embodiments may identify data chunks of the recent backup session present in the stored data blocks prior to the recent backup session.

Growing community of inventors

Beijing, China

Liqiu Song