A Distributed Graph Data Storage and Computing Framework

Graph has been widely adopted in Social Network Service (SNS) applications because it is very easy to represent user relationship information. In order to achieve high performance, prevalent algorithms tend to keep the graph data in memory to speed up the data access. However, failing to balance the costs of random access patterns and physical memory consumption hurts the effectiveness. In this paper, we propose a novel graph data processing framework to solve these issues. It develops a novel graph organization model to minimize the costs of graph data accesses and reduce the memory consumption. We compare our system with other graph processing systems on several widely used applications using large-scale real world SNS data, and results show that it outperforms the current mechanisms.