Research And Design Of Traffic Detection System

Since1969, the Internet has developed into the most popular media, and penetrated into allareas of modern society, becoming an indispensable part of human society. On the basis ofnetwork media and protocol, the Internet transfers video, audio, images, text and other data.Human society is enjoying a high level of service brought by modern information technology, atthe same time, with the extensive growth of network scale，access to various heterogeneousnetwork，the growth of network users, the traditional network structure can not adapt to the rapidchanges, Thus the supervision department of network has an urgent need to detect trafficinformation.The research object of the paper is network traffic detection. It has three steps for this results,At first, analysising the various network technologies of detection; Then, based on the analysis,putting forward a set of detection mechanism; At last, testing the sheme, the test to verify that thescheme can efficiently and accurately detect a variety of traffic, effectively managing the networkand providing differentiated services. The main achievements are4points:(1). An in-depth analysis of the current variety of network detection technology, comparingthe advantages and disadvantages of each technique, summarizing a variety of detectiontechnology strengths, putting forward to an innovative solutions of network detection.(2). Through the detailed study of various detection techniques, summarizing variousparameters of the detection system to a feature library, including a new formation. feature libraryhas the advantage of format rigorous, fast resolution, fast loading.(3). Putting forward a set of pre-treatment process for the detection. By studying the OSIseven-layer model of network protocol and the input information required for detection, secondarydetection processing mechanism is proposed to improve the detection rate and reduce systempower consumption.(4). Testing the detection system. Firstly, setting Google talk as an example, grabbing datasamples for analysis, writing signatures to the feature library, analyzing the testing results; Then,testing today’s popular network applications under different network structure, the test results showthat this detection system has wide detection range, high precision, strong feasibility.