The Study And Implementation Of Network Traffic Model Based On The Traffic Characteristics

The network traffic model is the basis of traffic prediction and network performance evaluation,it not only can well reflect the characteristics of real network traffic but also can well guide the practical application about network traffic.With the rapid development of network technology,the number of network node grows with exponent,the network applications also become diversified and complicated,the network protocols have also been gradually increased and changed,people do research on network traffic model deeper and deeper.Network traffic model based on traffic characteristics not only provides authentic and reasonable analysis for optimizing network structure,improving network protocols and network service quality,but also can provides authentic test environment and conditions for testing network equipment and security products,building networks system.It first analyzes and summarizes the existing network traffic model.Summarize the models from the Poisson model in early 1970s and 1980s to the self-similar model and multi-fractal model,then later developed various forecasting models,meanwhile analyze their strength and weakness.Then designs and implements the network traffic model system based on traffic characteristics.The system first captures high-speed packets,then extracts and analyzes the traffic characteristics,finally calculates the probability density function and cumulative distribution function of the traffic characteristics,using the maximum correlation coefficient to determine the distribution model of traffic characteristics.It is designed and implemented a system of network traffic model based on traffic characteristics.It first captures high-speed network packets from real network.Then network packets are distributed to multiple threads which use to process packets and analyze packets,then extracts and counts traffic characteristics such as five tuple,packet size,packet number,fragmented packets number,traffic bandwidth,and so on.Finally we draw the probability density curve and cumulative distribution curve of this traffic characteristics,and calculate the correlation c.oefficient with common distributions(such as Pareto,Normal,Lognormal,Exponent,Poisson,Weibull).We find that the lognormal distribution can be used to describe the packets number,fragmented packets number,traffic bandwidth,packet size and flow number for real network.