Compressing Streaming Graph Data Based on Triangulation

Liang Zhang,Ming Gao,Weining Qian,Aoying Zhou

DOI: https://doi.org/10.1007/978-3-319-45835-9_15

2016-01-01

Abstract:There is a wide diversity of applications for graph compression in web data management, scientific data processing, and social data analysis. In real-life applications like social media data processing, elements in a graph, typically vertices and edges, are arriving continuously. Compressing the graph before storing it in a database is important for real-time processing and analysis, while being a challenging yet interesting problem. A streaming lossless compression method, named as STT (streaming timeliness triangulation), is introduced in this paper. It is a time-efficient method for compressing a streaming graph, which differs itself from static graph compression methods in that: (1) it’s able to compress streaming graph without occupying extra storage; (2) it can achieve both low compression ratio and high throughput over the streaming graph; (3) it supports efficient graph query processing directly over compressed graphs. Thus, it can support a wide range of streaming graph processing tasks. Empirical study over a paper co-author graph and a real-life large-scale social network graph has shown the superiority of the newly proposed method over existing static graph compression methods.