Flow Cluster Algorithm Based on Improved K-means Method
Shi Dong, Dingding Zhou, Wei Ding, Jian Gong · IETE Journal of Research · 2013
AbstractTraffic classification is currently an important challenge for network management. In recent years, some traffic classification and identification algorithms have been proposed; identifying encrypted application traffic represents an important issue for many network tasks including quality of service. Port number-based classifiers work only for well-known applications and signature-based classifiers are not suitable for encrypted packet payloads. So researchers tend to identify network traffic based on behaviors observed in network application. But the results are so far limited in scope and frequently disappointing. In this paper, flow identification method is proposed to identify network flows based on traffic statistic, which adopt improved k-means cluster algorithm (SA-k-means) to classify traffic, and analyze the impact factor of cluster. Also, experiment results show SA-k-means method is effective.