Research And Implementation Of Network Traffic Simulation Technology Based On Traffic Characteristics

Network is an important infrastructure of today's information society.With the rapid development and popularization of the network,the network structure and size have become ever more complex and enormous.For the sake of performance optimization of the network and network devices,network traffic simulation has played an important role and got more and more attention.While domestic simulation technology is not mature enough for test equipment.Therefore,to analyze and study network traffic simulation has important significance.This thesis analyzes the self-similarity and fractal nature of the network traffic,sums up the background knowledge and related research in the field of network traffic generation.By analyzing the advantages and disadvantages of the generation method we provide a multi-angle and comprehensive contrast of the related traffic simulation technology.On this basis,a new simulation algorithm based on package level traffic feature has been proposed.One can set parameters such as package distribution,protocol distribution and so on flexibly so as to simulate original traffic.Meanwhile,based on the needs of data compression of the traffic feature,an improved algorithm based on Douglas-Peucker has been put forward.The algorithm reduces the storage of data points to improve processing efficiency.Meanwhile,on the basis of these theoretical analysis and research,we design the traffic characteristics simulation system.The system can construct traffic that has the same characteristic distribution of the original flow.The system has good man-machine interaction,generation task and capture task can be easily configured,and can display characteristics of the flow from different scales.The system is good performed,fully functionally and has been applied to the actual test environment.