Research On Simulation Technologies And Algorithms For Packet-switched Network

Recently, the communication network technologies have developped rapidly and the packet-switched network targetting for huge-capacity, high-speed and multi-service has gained more and more attention. This leads to the increasing importance of research and development on all kinds of new techniques. Network Simulation based on discrete event simulation technology is not only an important means of network planning and design, but also a key approach to new technology verification and improvement. Currently all the popular simulation software suffers from such shortcomings as low-usability and high-cost. Especially, they fail to statisfy the demand of high-strength simulation for the future network. Therefore, it is of great significance to design an easy-to-use simulation software meeting the market requirements. QoS is the key to the packet-switched network and should be considered with high priority for the network simulation. The thesis focuses on the packet switched network simulation system and its performance simulation solution, proposing a feasible solution for high-strength simulation.Firstly, the thesis gives an introduction to the packet-switched network with its QoS techniques and discusses discrete event system simulation technology with some common-used simulation software. Secondly, according to the requirements, the thesis brings forward the overall architecture of packet switched network simulation system followed by detailed description of the structure and functions of simulation computing subsystem. Two key algorithms, statistcal link utilization ratio algorithm and service simulation terminating algorithm, are proposed for the service simulation. Thirdly, to enhance the simulation accuracy, the thesis offers the idea of precise modeling based router design and proposes a three-level configuration scheme, covering device-level, board-level and unit-level. The concrete design sheme and test results are worked out for the QoS functional units. Finally, the thesis proposes an improved event scheduling algorithm with the help of CPU cache. The testing results have shown that the algorithm can effectively raise the system efficiency with regard to high-strength event scheduling.