Resarch And Application Of Traffic Identification Based On The Hybrid Behavior Characterstics

With the rapid development of Internet technology and the arrival of information age, Internet applications have been deeply affected every aspect of human social life. Use of the Internet caused the explosive growth of network data, thus increased the burden of network security and network monitoring task. Traffic identification is the premise of network security and network analysis, so it is very import to classify the traffic at the entrance of the network. The accurate recognition of network traffic not only insight into the entire network status, but also can for specific for precise regulation, then it can ensure network speed, while to prevent network behavior, such as network attack or network storm.However, with the increasing number of business requirements, a variety of network applications get rapid development. This kind of situation has brought serious challenges to existing traffic classification methods. The traditional traffic classification technology has played an enormous role in the different stages of the development of the network, but there are different technical bottlenecks in the complex network.This thesis draws lessons from the experience of traditional traffic identification technology, and from the view of the graph, a new traffic identification model is proposed, which is combined with the popular detection and complex network. Its work is divided into the following parts. First of all, based on the statistical characteristics of the traffic flow,the classification method of machine learning is applied in the field of traffic identification, to abstract the behavior characteristics of the flow in the process of communication, and select the appropriate behavior characteristics. Secondly, the complex network is studied, and its main idea and algorithm are analyzed deeply, while the application of graph feature in the field of traffic identification is proposed. Then, through combining flow behavior and graph connection behavior, the research on how to construct a more comprehensive model of mixed traffic identification. Finally, we use the big data platform Spark to achieve efficient and accurate flow identification system. The new traffic identification model is a powerful supplement to the classification model which is only concerned with a single level traffic feature,and can achieve better classification results. With the advantage of Spark in large data processing, the traffic identification system is more suitable for the processing of massive data traffic.