【作者】 钟伯成； 韩江洪； 杜兆芳； 张杰；

【Author】 ZHONG Bo-Cheng1,2+ HAN Jiang-Hong1 DU Zhao-Fang1 ZHANG Jie11(School of Computer & Information, Hefei University of Technology, Hefei, 230009,China)2(Department of Computer Science & Technology, University of Hefei, Hefei, 230022,China)

【机构】 合肥工业大学计算机与信息学院；

【摘要】 提出一种新的基于最优化的分配多媒体多速率组播流速率的方法。方法权衡了各流的QoS、异构性及公平性等,仿真实验证明了算法的有效性。更多还原

【Abstract】 Multimedia data transmission over the Internet is a challenging task. When streaming multimedia data over the network, multicast has the advantage of sending only a single copy of the data across the network while preserving the network bandwidth. Multirate multicast allows the receivers of the same multicast group to receive service at different rates, which is therefore more desirable for multimedia applications with heterogeneity in a large network. Although multicast has long been viewed as an attractive service for multimedia applications in the Internet and has become an important research area, there has been little multicast deployment in the Internet. One main impediment to the wide deployment of multicast is the existing service model’s inability to provide an effective rate control mechanism. Indeed, rate control for multirate multimedia multicast faces many challenges such as heterogeneity, fairness, real-time constraints, and quality of service etc. In this paper, we present a novel rate control scheme for multirate multimedia multicast that tries to meet those properties required. In the first place, we, based on microeconomic theory, convert the problem of end-to-end rate control into an optimization problem, which maximize the aggregate utility. The objective is to allocate rates for all sessions in a network to satisfy inter-session fairness. We propose a distributed and scalable rate allocation algorithm, which has well-defined fairness properties and allows for the possibility of providing user priorities through the proper setting of certain design parameters associated with a particular user or application type. In the second place, with the result of rates allocation above, we cast the joint problem of layering and rate allocation in a multicast session into an optimization of total system utility among all the users of a group and derive an efficient algorithm to solve the optimal layering and rate allocation problem. Our rate control approach achieves a good tradeoff between user’s satisfaction and network efficiency and meets both the inter-session fairness and intra-session fairness. Simulation results are given to demonstrate that our method improves system performance.更多还原