J4 ›› 2013, Vol. 40 ›› Issue (3): 160-165.doi: 10.3969/j.issn.1001-2400.2013.03.024

• Original Articles • Previous Articles     Next Articles

Improved distributed traffic sampling algorithm  for high speed network

SUN Yu1;JIANG Fuwei1;XIA Jingbo1;ZHAO Xiaohuan1;SHEN Jian1;LI Minghui2   

  1. (1. School of Information and Navigation, Air Force Engineering Univ., Xi'an  710077, China;
    2. Air Force Logistics Department, Beijing  100720, China)
  • Received:2012-09-27 Online:2013-06-20 Published:2013-07-29
  • Contact: SUN Yu E-mail:458657673@qq.com

Abstract:

The sampling algorithm based on mask matching is a distributed sampling measurement technology which can be used in a high speed network. But there exists a problem that some masks used in the algorithm result in a large measurement error. In order to solve the problem, the paper proposes an improved scheme that aims to enhance the randomness of its sample. The idea of the scheme is to put the identification field into xor operation to get a new identity value before mask matching and sampling. Firstly, the paper demonstrates theoretically that the xor operation can enhance its computed result's randomness effectively. Then it is checked statistically with the realistic traffic data. Finally, the improved algorithm is tested from two aspects of effectiveness and stability. Experimental result shows that the evaluation indexes of the improved algorithm are better than those of the original algorithm. Hence, any sampling mask can be used in the improved algorithm.

Key words: distributed sampling, mask matching, identification field, randomness, xor operation