Traffic Classification in Mobile IP Network

Traffic classification is an essential task for network management. Thus many researchers have paid attention to initial sub-flow features based classifiers for achieving online traffic classification. However, the existing classifiers cannot classify traffic effectively in mobile IP network, because they cannot always capture the initial sub-flow when a communicating node moves and its point of attachment changes. Desirable classifier should thus be capable of traffic classification based on not only initial sub-flow but also various types of sub-flow. In this paper, we propose sub-flow selection with application behaviors, and the method solves a critical problem of how to select appropriate sub-flows for achieving traffic classification in mobile IP network. Although at a very early stage of development, the proposed method shows promising preliminary results through the experiments on a reduced set of applications.