US 12,271,357 B2
Methods and systems for scalable deduplication
Yong Yang, Beijing (CN); Xianbo Zhang, Plymouth, MN (US); Weibao Wu, Vadnais Heights, MN (US); Chao Lei, Beijing (CN); Yafeng Wang, Beijing (CN); Haigang Wang, Shoreview, MN (US); and Lulu Wei, Beijing (CN)
Assigned to Cohesity Inc., San Jose, CA (US)
Filed by Veritas Technologies LLC, Santa Clara, CA (US)
Filed on Jul. 5, 2023, as Appl. No. 18/347,395.
Application 18/347,395 is a continuation of application No. 16/698,288, filed on Nov. 27, 2019, granted, now 11,741,060.
Prior Publication US 2023/0350863 A1, Nov. 2, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 16/215 (2019.01)
CPC G06F 16/215 (2019.01) 21 Claims
OG exemplary drawing
 
1. A method comprising:
in response to receiving a search request from a client computing system, determining whether a subunit of data has been stored as part of a backup image, wherein
the search request comprises a signature that identifies the subunit of data,
the search request is received at an assigned node that is one of a plurality of nodes of a cluster,
one or more nodes of the plurality of nodes are identified in a catalog as storing one or more corresponding portions of the backup image, and
the determining comprises
performing a lookup operation, wherein
the lookup operation is performed by the assigned node,
the lookup operation causes the assigned node to query a reference database, using the signature,
the reference database comprises a first plurality of references,
one or more of the first plurality of references refer to a first plurality of signatures that identify a first plurality of subunits of data of a portion of the one or more corresponding portions, and
the assigned node is identified in the catalog as managing the portion, and
in response to the signature not being found in the first plurality of signatures,
forwarding the search request to a remote node, wherein
 the remote node is a one of the plurality of nodes, other than the assigned node,
 the remote node is identified in the catalog as managing another portion of the one or more corresponding portions,
 the another portion is a one of the corresponding portions that is other than the portion,
 the forwarding the search request to the remote node causes the remote node to perform another lookup operation that causes the remote node to query another reference database, using the signature,
 the another reference database comprises a second plurality of references,
 one or more of the second plurality of references refer to a second plurality of signatures that identify a second plurality of subunits of data of the another portion, and
in response to the assigned node receiving an indication from the remote node that the signature was not found in the second plurality of signatures, indicating that the subunit of data should be processed as a unique subunit of data, wherein
the cluster implements a shared-nothing architecture at least by virtue of the assigned node managing the portion and the remote node managing the another portion.