Research On Technologies Of Multi-patterns Network Traffic Generation

As the basis of network congestion control, network bandwidth allocation and network performance evaluation, network traffic generation technology provides technical support for development of network technology. As the size of the network continued to expand, the network structure is becoming more complex, the composition of network traffic is becoming more diverse, and the characteristics of network traffic are increasingly complex. The existing network traffic generation models are non-uniform, the type of traffic generated by current network generation tool is singleness and performance is not high enough. In face of these problems, this thesis discusses a multi-patterns network traffic generation technology and researches its applications.The main contents of the thesis include:1. The existing network traffic characteristics are analyzed in this thesis. By researching a variety of network traffic generation models, the thesis compares each model from theory to simulation, and provides much range of research for multi-patterns network traffic generation model.2. The wavelet method is introduced to self-similar network traffic generation technology. By improving the random midpoint displacement method, a solution for self-similar network traffic generation is presented.3. The random process theory is introduced in combination of wavelet method. By changing the variables between wavelet coefficient and scale coefficient, a new stochastic network traffic generation technology is presented to achieve a variety of stochastic network traffic generation on destined probability distribution, destined expectation and variance.4. A unified description of multi-patterns network traffic generation is used for combining self-similar network traffic generation technology with stochastic network traffic generation technology. In face of the non-uniform characteristics of network traffic, a multi-patterns network traffic generation technology is proposed by the connection of inter-frame delay, Hurst and types of network traffic.5. The multi-patterns network traffic generation technology is applied to multi-patterns network traffic generation system by using multi-core layered architecture network processor as a hardware development platform. By putting the system into applications of network attack behavior detection, network bandwidth measurement as well as network testing, the test results prove the effectiveness of the above technologies.