US 12,461,969 B2
Graph data partitioning
Wei Qin, Hangzhou (CN); Jiping Yu, Hangzhou (CN); Xiaowei Zhu, Hangzhou (CN); and Wenguang Chen, Hangzhou (CN)
Assigned to Alipay (Hangzhou) Information Technology Co., Ltd., Hangzhou (CN)
Filed by ALIPAY (HANGZHOU) INFORMATION TECHNOLOGY CO., LTD., Zhejiang (CN)
Filed on Dec. 22, 2023, as Appl. No. 18/394,497.
Application 18/394,497 is a continuation of application No. PCT/CN2022/131042, filed on Nov. 10, 2022.
Claims priority of application No. 202111345319.4 (CN), filed on Nov. 15, 2021.
Prior Publication US 2024/0143657 A1, May 2, 2024
Int. Cl. G06F 16/901 (2019.01); G06F 11/34 (2006.01)
CPC G06F 16/9024 (2019.01) [G06F 11/3433 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method for graph data partitioning, comprising:
performing a layer-1 partitioning on graph data to generate a first plurality of datasets, wherein each dataset in the first plurality of datasets corresponds to a respective node in a plurality of nodes in a distributed cluster, and wherein performing the layer-1 partitioning comprises:
partitioning vertices in the graph data into the first plurality of datasets;
partitioning edges in the graph data into the first plurality of datasets that include target vertices of the edges,
performing a layer-2 partitioning on the first plurality of datasets to generate, for each node in the plurality of nodes in the distributed cluster, a second plurality of datasets, wherein for each node, a number of the second plurality of datasets is dependent on a count of threads or a count of processes of the node; and
performing graph computation on the second plurality of datasets by each node in the plurality of nodes in the distributed cluster.