Research And Implementation Of Protocol Identification And Traffic Classification On Application Layer

With the rapid development and widespread popularity of Internet technology, network operation and management of is facing more and more issues. In recent years, the number of Internet users and network bandwidth of the international gateway has maintained increasing growth. With the scale of the Internet expanding for the increase of bandwidth and user scale, variety of network applications have emerged and unknown protocols increase constantly, which making the network more difficult to manage. With the development of mobile Internet, cloud computing and other emerging technologies, we can foresee that type of applications will be more and more thriving. A sharp increase in the type and number of network applications has greatly promoted the growth in network traffic. Therefore, if network traffic is not effectively managed, normal business will not properly transmitted. In order to manage Internet bandwidth effectively, find a reasonable solution to network congestion and ensure the QoS of the major transmission of critical business, application layer protocol identification technology is always a key issue of research.On one hand, with the extensive use of cryptographic protocols, more and more application-layer protocol uses encryption protocol for data security and information privacy. The existing protocol identification technology has encountered a bottleneck in terms of identifying encrypted traffic. On the other hand, in order to provide better service and user experience, plenty of applications such as Tencent QQ integrate multi-application and multi-protocol. Therefore, there calls for a fine-grained technology for protocol identification and traffic classification. Towards this issue, especially for the issue of encrypted protocol, we have a research on the application layer protocol, aiming at presenting effective method and technique.This thesis describes the purpose and significance of the techniques of protocol identification and classification on application layer, analyzed the main methods used to solve the problems. Towards the challenges encountered by now, we have proposed methods of encrypted protocol identification and fine-grained classification through long-term observation, analysis and experiment. Then we design the overall framework and structure of the system of identification and traffic classification on application layer. By the test of encrypted traffic fine-grained classification, performance, stress and comparison, we verify the reasonableness of the system architecture and validate effectiveness of the method. And then we achieve a good effect.