The Research Of Session Initiation Protocol Based On Generalized Stochastic Petri Net

Considering the promising foreground application of Session Initiation Protocol(SIP)in network and communication field,the research of SIP should be wholly designed,and made formal description to the implementation of SIP. Based on SIP's formal model , its property and performance can be conveniently analysed and evaluated for properly designing, developing and optimizing the protocol. The task is of great academic and applied value to solve the problem of SIP's application.Generalized stochastic Petri Ne(tGSPN)is a kind of diagrammatic language, and has strong power of describing and analyzing complicated systems. Not only a generalized stochastic Petri Net-Based technology of establishing and analysing the protocol model can be used to comprehend the system performance in service process of protocol and resource scheme,but also it can give a method of making the system implement optimize and modify the logic structure of system.At present, the research of SIP mainly focuses on how to enhance the performance of SIP-based service network and how to improve QoS of SIP's component. However, this dissertation emphasizes mainly upon SIP in itself for enhancing the performance of protocol. On the basis of the abundant research of signaling interchange of SIP,stochastic Petri net(SPN) model of SIP is established. The fundamental property of SPN,its mathematical theory and analyse technique is given to verify the feature of the protocol.The dissertation focuses on mathematical analyse technique and simulation analyse for SIP. Markov chain of the model was constructed in terms of GSPN model of SIP. Through the abundant calculation about markov chain,performance index of SIP is analysed and studied. Moreover,basing on SPN model of SIP,the bottleneck of system is analysied. Basing on the simplification of GSPN model,upper limit of throughput is studied. Combining model-based analysis and software simulation,quality of service of SIP is analysied and improved.The main innovative results of the dissertation are as follows:(1)Studying signaling interchange of SIP, stochastic Petri net model of SIP is established. The property of the protocol is verified and analysed by using mathematical theory of SPN, which can offer a approach for evaluating and improving the performance of SIP. (2)By GSPN model-based analysis of SIP and software simulation quality of service of SIP is simulated and studied. Performance analysis and simulation results show that the optimal size of the buffer in the place P2 can contribute to reducing call setup time and improve the throughput of system, which gain a tradeoff between the performance and the number of buffer.(3)Using mathematical theory and analyse technique of GSPN performance index of SIP is studied and evaluated.