Journal of Beijing University of Posts and Telecommunications

  • EI核心期刊

Journal of Beijing University of Posts and Telecommunications ›› 2020, Vol. 43 ›› Issue (1): 97-103.doi: 10.13190/j.jbupt.2019-076

• Reports • Previous Articles     Next Articles

A Heavy Hitter Detection Mechanism in Software Defined Networks

XING Chang-you1, LI Dong-yang1, XIE Sheng-xu1, ZHANG Guo-min1, WEI Wei2   

  1. 1. Army Engineering University, Command and Control Engineering College, Nanjing 210001, China;
    2. Corps 31106, Nanjing 210016, China
  • Received:2019-05-11 Online:2020-02-28 Published:2020-03-27
  • Supported by:
     

Abstract: SampleFlow, a heavy hitter detection mechanism in software defined networks, is proposed to solve the problems of low detection accuracy and high measurement cost. By combining the technical advantage of sFlow and OpenFlow, SampleFlow firstly detects a set of suspicious heavy hitters by using the coarse-grained sFlow sampling method, and then installs measurement flow entries on specific OpenFlow switches to perform a fine-grained measurement on these suspicious heavy hitters, so as to determine the true heavy hitters. Besides, SampleFlow also uses a sampling position optimization method to decrease the sampling redundancy. Experiment results show that SampleFlow can decrease the measurement cost, and increase the heavy hitter detection accuracy effectively.

Key words: software defined network, network measurement, heavy hitter detection, sampling

CLC Number: