PARALLEL GRAPH PATTERN MATCHING IN MASSIVE NETWORKS BASED ON MAPREDUCE

In many popular applications, e.g., social network and bioinformatics, there is a growing need for supporting efficient graph pattern matching over massive graph data. Since the problem of graph pattern matching is NP-complete, it is challenging to operate over large graphs. Existing solutions to this problem are centralized, thus have limited use in practice. In this paper, we propose a parallel approach for graph pattern matching based on the popular computing framework Map Reduce. The proposed approach divides query and data graphs into paths, and then performs pattern matching in a path-at-a-time manner. Our experiments on real datasets verify its efficiency.