QoS Performance Evaluation Of Self-Similar Netweok Based On ETS

With the rapid development of high-speed network and multimedia technologies, more and more multimedia communication service have been proposed. Traditional internet is designed for non-real-time data communication service, it is unable to meet the high quality of service requirements of a variety of multimedia applications, therefore, the control and assurance of QoS based on user requirements is a very common need, it is also a challenge for the internet. In addition, after research and analysis on a large of number of network applications, it indicate that self-similar is common in the actual network applications, the feature of self-similar is different from short-range dependence, therefore, the conclusion of flow control for traditional model is not suitable, and a model is need to describe the feature of self-similar. In this paper, we combine traffic shaping mechanism and traffic scheduling mechanism, the fractal regulator is used to shape traffic, the traffic output envelope is treated as the arrive curve of the ETS system, and a new service curve based on ETS scheduling algorithm is proposed. Next, the arrive curve and service curve is used to establish a self-similar network QoS performance evaluation model based on network calculus theory, the determine upper bound and the statistical upper bound of the self-similar network QoS performance based on ETS is deduced, finally, the upper bound of the QoS performance is proved to be effective by experimental analysis, the results is significant for providing high-quality QoS guarantee with self-similar network.