Improved processing speed of traffic classification based on payload signature hierarchy

The rapid development of the Internet has necessitated effective network management. Traffic classification is the process or protocol used to manage network traffic. Signature-based classification is the best known traffic analysis technique. This technique, however, has the negative aspect that processing speeds slow as the number of signatures increases. In this paper, we propose a novel method that elevates the processing speed by using a signature hierarchy. In addition, we propose a representative signature concept that reorganizes the conventional one-dimensional signature methodology. Signatures can be structured through representative signatures in a hierarchical structure. The traffic classification system can then classify traffic based on structured signatures. Using this new matching algorithm, traffic can be classified faster than with conventional methods. The feasibility of our proposed system has been demonstrated with experiments on campus traffic data.