Third International Symposium on Parallel Architectures, Algorithms and Programming (PAAP 2010) (2010)
Dec. 18, 2010 to Dec. 20, 2010
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/PAAP.2010.69
Peer-to-Peer technology is one of the most popular techniques nowadays, and it brings some security issues, so the recognition and management of P2P applications on the internet is becoming much more important. The selection of protocol attributes is significant to the problem of P2P identification. To overcome the shortcomings of current methods, a new P2P identification algorithm based on genetic algorithm and particle swarm optimization is proposed. The attributes of network traffic flows are selected and assigned the corresponding weightings according to their importance by evolutionary algorithms. The experimental results show that this algorithm can effectively select the subset from multiple attributes that can best reflect the differences among some most popular P2P protocols and also between P2P and non-P2P protocols. The identification rate is improved by the method of feature weighting calculated by particle swarm optimization. With this algorithm, the average identification rate of popular P2P protocols reaches to 96.3%.
Classification algorithms, Protocols, Particle swarm optimization, Nickel, Clustering algorithms, Algorithm design and analysis, Machine learning algorithms
J. Tan, X. Chen and M. Du, "A Novel P2P Identification Algorithm Based on Genetic Algorithm and Particle Swarm Optimization," Parallel Architectures, Algorithms and Programming, International Symposium on(PAAP), Dalian, Liaoning China, 2011, pp. 22-29.