Towards Bipartite Graph Data Management

摘要

Bipartite graphs are widely used to model social networks and web data. However bipartite graph data management (BGDM for short) is not fully studied. Without appropriate indexing structures, query and analysis operations on bipartite graphs are not efficient. In this paper, we raise the issues of BGDM and present architecture of BGDM. Then we propose a logic graph structure (i.e., star) for indexing bipartite graph to improve common operations efficiently. Furthermore, we propose a star-based block structure to store bipartite graphs and the relevant query algorithm. The star-based method may avoid loading the whole block for vertex queries using Bloom filter. Finally, our experiments show that the block design is effective and feasible for vertex query algorithm.