Efficient Parallel Cohesive Subgraph Detection

Abstract

Community detection is a fundamental graph analytic task. However, due to the high computation complexity, many community detection algorithms cannot handle large graphs. In this chapter, we investigate a special community detection problem, that is, cohesive subgraph detection. Here the target cohesive subgraph is k-truss, which is motivated by a natural observation of social cohesion. We propose a novel parallel and efficient truss detection algorithm, called PeTa. PeTa produces a triangle complete subgraph (TC-subgraph) for every computing node. Based on the TC-subgraphs, it can detect the local k-truss in parallel within a few iterations. We theoretically prove, within this new paradigm, the communication cost of PeTa is bounded by three times of the number of triangles, the total computation complexity of PeTa is the same order as the best known serial algorithm, and the number of iterations for a given partition scheme is minimized as well. Furthermore, we present a subgraph-oriented model to efficiently express PeTa in parallel graph computing systems. The results of comprehensive experiments demonstrate, compared with the existing solutions, PeTa saves 2× to 19× in communication cost, reduces 80% to 95% number of iterations, and improves the overall performance by 80% across various real-world graphs.