Real-time internet traffic identification based on decision tree

Real-time Internet traffic identification is always a hot research topic in recent years. It involves quality of service, network accounting, Intrusion Detection and so on. Traditional identification approaches, such as those based on port and payload analysis, are no longer applicable in actual networks. In this paper we present a machine-learning approach, independent of port numbers, to accurately classify Internet traffic using decision tree. In our work, we think over not only the accuracy, but also the time cost. We use FCBF to remove redundant features and C4.5 algorithm to build the classification model and guarantee both accuracy and efficiency.