VSEP: A distributed algorithm for graph edge partitioning

Abstract

With the exponential growth of graph structured data in recent years, parallel distributed techniques play an increasingly important role in processing large-scale graphs. Since strong connections exist between vertices in graph data, the high communication cost for transforming boundary data is unavoidable in the distributed techniques. How to partition a large graph into several partitions with low coupling and balanced scale becomes a critical problem. Most of research in the literature studies vertex partitioning methods, which leads us to reconsider an alternative approach for edge partitioning. In this paper, we propose a distributed algorithm for graph partition based on edge partitioning, named as VSEP. A novel vertex permutation method is used to partition the large graphs iteratively. Experimental results indicate that VSEP reduces the number of times vertices are cut by about 10%∼20% comparing with a state-of-the-art algorithm while retains the scale balance.