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.
CITATION STYLE
Zhang, Y., Liu, Y., Yu, J., Liu, P., & Guo, L. (2015). VSEP: A distributed algorithm for graph edge partitioning. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9532, pp. 71–84). Springer Verlag. https://doi.org/10.1007/978-3-319-27161-3_7
Mendeley helps you to discover research relevant for your work.