Study On Background Traffic Model In Network Simulation

Network simulation is a key method to do research on network behavior and to analyze the network protocols; it has a great value in science research and application. Network simulation behavior mainly observed network impact on network security incidents through analyzing the data of traffic generated by the simulation. The traffic generated by simulation consists of application traffic generated by the network security incidents and background traffic existed in the actual network. A reasonable and realistic traffic model can provide a reliable traffic for network simulation behavior, and it can improve the realism of the simulation of network behavior. Traffic model has important research value and significance, this paper aims to undertake research for this.This dissertation takes traffic model as research object, analyzes and designs three kinds of traffic model. It designs and realizes the applications of traffic flow generator corresponded with the three kinds of traffic model in NS2. There are mainly four parts in this dissertation, as follows:To solve the problem about some traffic models only reflect some characters of real traffic, but not the state of a time, this dissertation presents a model method to simulate traffic, according to real traffic collected. By analyzing packets arriving procedure, we consider that the number of arriving packets at fixed time period follows Poisson distribution and the packet arrival interval follows Exponential distribution. Provide for the simulation of network behaviors, by designing traffic generator related to Poisson model and adding it to NS2. In the research process of ON/OFF model based on Pareto distribution, we find some deficiencies about the model in NS2 about the setting of heavy tail parameter and the calculation of minimum value of OFF. So this dissertation improves the ON/OFF model in NS2 and redesigns traffic generator in NS2. Taking into account the simulation expense is too large in the adding process of background traffic application, this dissertation designs and realizes an algorithm to simplify addition of traffic application.Finally, through the three kinds of background traffic model and background traffic adding algorithm experiments, which showed that three kinds of background traffic model meet the design features and its corresponding traffic generator application can provide the background traffic for the simulation of network traffic behavior; background traffic adding algorithm can reduce the number of adding applications, enhance the performance of the network behavior modeling and optimizing the implementation of the background traffic model.