Research On Traffic Identification And Monitoring System Based On DPI

History of Internet development for nearly 50 years,not only brought great changes for the field of science and technology,but also to promote the development of the whole world.From the traditional web browsing to present complex network applications,people have become increasingly dependent on the network to innovation everything.Internet industry has also been a qualitative leap,this qualitative change in the performance of high-speed broadband Internet market,quality service,accurate demand information and safe environment.To ensure a secure network environment,this thesis designed a DPI-based network traffic identification and monitoring system for monitoring network traffic.According to the functions of the system module division,outlined various types of network traffic transmission technology,specifically summarizes the technical characteristics of P2 P networks.Analysis of the detection method port,network layer,application layer,transport layer detection technology based on the introduction of DFI for encrypted traffic detection mechanisms to improve the recognition of the internet traffic.By analyzing the DPI-based detection algorithm,an improved algorithm BM-AC,and with examples Baidu cloud disk analysis matching process.Ensuring detection speed while reducing the probability of occurrence of erroneous recognition and identification of leakage,while simplifying the rules described in the knowledge base to write complicated string.DFI part in detecting neural network algorithm and statistical characteristics,structural characteristics matching model to achieve recognition on the part of the encrypted traffic.Finally,the system testing phase,select the part of the application for traffic identification and blocking tests.Recognition rate of 95%,blocking rate of 100%,to verify the reliability of the program.