Article :Browse 1480 Download 1312
Received:April 02, 2014 Revised:September 09, 2014
Received:April 02, 2014 Revised:September 09, 2014
Abstract:Detection of super points is very important for some network applications such as network security and network management. Due to the imbalance between the massive amount of traffic data in high-speed networks and the limited system resource, it is a great challenge to accurately monitor traffic in high-speed networks online. With the development of multi-core processers, the parallelism of multi-core processers becomes an effective method of improving performance. Consider that the existing method based on flow sampling has heavy computing load, low detection accuracy and poor timeliness, this article presents a new approach, PDS (parallel data streaming). This method constructs parallel reversible sketch and builds a compact summary of node connection degrees. Addresses of super points are reconstructed by a simple calculation without address information of nodes. This makes PDS efficient and precise. The experimental results show that PDS is superior to the CSE (compact spread estimator) and JM (joint data streaming and sampling method) methods. Consequently, the proposed method satisfies application requirement of traffic monitoring in high-speed networks.
Foundation items:National High Technology Research and Development Program of China (863) (2015AA015603); Jiangsu Future Networks Innovation Institute: Prospective Research Project on Future Networks (BY2013095-5-03); Six Talent Peaks of High Level Talents Project of Jiangsu Province (2011-DZ024); Research and Innovation Program for College Graduates of Jiangsu Province (KYLX_ 0141)
Reference text:
ZHOU Ai-Ping,CHENG Guang,GUO Xiao-Jun,LIANG Yi-Xin.Parallel Data Streaming Method for Detection of Super Points in High-Speed Networks.Journal of Software,2016,27(7):1841-1860
ZHOU Ai-Ping,CHENG Guang,GUO Xiao-Jun,LIANG Yi-Xin.Parallel Data Streaming Method for Detection of Super Points in High-Speed Networks.Journal of Software,2016,27(7):1841-1860