Design And Implementation Of Network Traffic Monitoring System

With the network application technology becoming more and more popular, there are some problems that the number of bandwidth requirements is increased by a binary escalation and the bandwidth resources become increasingly strained. It is important to design a network traffic monitoring system to control network traffic. The main jobs are included in the following aspects:Through the study of many kinds of traffic identification technologies, the network traffic identification technology based on application layer is focused on. There are two load characteristics which are summarized based on application layer: fixed load and fixed offset,fixed load and variable offset. A recognition scheme is put forward according to the different characteristics. A library is designed by using Aho-Corasick automaton algorithm(AC) and regular expression algorithm to write characteristics of rules, which not only ensure the accuracy of identification protocol, but also improve the efficiency of matching engines.With researching the efficient algorithm of finite automaton, a matching algorithm combining AC algorithm with regular expression algorithm is selected for the design of the engine. There is a problem that the states of deterministic finite automaton(DFA) are expanded when compiling regular expression. An optimization of DFA grouping algorithm is put forward. The experiment results show that the optimization of DFA grouping algorithm can effectively reduce the number of DFA states, thus decreasing the storage spaces.A traffic monitoring system which includes several modules is designed. A method of writing and compiling rules in library modules is presented. In traffic identification module the process of realization and the matching method of engine are described in detail.The experimental results of system's traffic identification and control are given. The performance of system is analyzed in detail, including throughput, stability, etc.