Researches On Strategies Of Multirate Multicast

Internet is a heterogeneous collection of networks and end-systems. Receivers may possessdifferent capabilities or be connected to the network through a variety of different accessspeeds. Under this circumstance, if only a single transmission rate is used in multicast (theusual mechanism), the conflicting requirements of a set of heterogeneous receivers cannot besatisfied simultaneously because of the simple reason that one size does not fit all. We refer tothis as intra-session fairness problem. The large scale and complexity of today's Internet makethings even worse. Multirate multicast is recommended as an ideal solution to this problem. Insuch protocols, data is delivered in multiple layers over separate multicast groups, receiversjoin as many layers as they can handle, and receive data at rates commensurate with theircapacities and available bandwidth. Even so, there are still some problems to be solved inmulticast, such as inter-session fairness problem and the poor performance of the end-to-endtraffic control strategies etc.In this paper, we focus on the research of layered multirate multicast strategies, aiming atsolving the problems of inter-session fairness and traffic control mechanism that blockmultirate multicast's advance. In order to get a traffic control mechanism that satisfies theMax-Min fairness requirements, we make some improvements and optimizations in thecustomary implementations of multirate multicast. Finally, we devise a Max-Min fairnessguaranteed layered multicast scheme in which routers participate in traffic control bymonitoring the available bandwidth and marking passing packets to explicitly indicate thecongestion, we name it FLM.Chapter 2 gives a discussion on the principle and research developments of layeredmultirate multicast. Then we point out the existing problems in it, and in Chapter 3, throughsome detailed examples of bandwidth sharing problems illustrated with simulations, weanalyze the necessity for layered multicast to ensure fair resource sharing among competingmulticast flows and TCP flows. So we propose a computation algorithm for Max-Min fair rateallocation which can be applied in the fair resource allocation in heterogeneous networkenvironment. As a complement, we also design an algorithm to discretely regulate the obtainedrates set in different granularity with the adoption of Bisection Method. Thereby we provideguarantee for the effective implementation of layered multirate multicast. This Max-Min fairrate allocation algorithm is an important component of the traffic control scheme in ourlayered multirate multicast scheme FLM.A router-assisted strategy of traffic control for FLM is devised in Chapter 4. In this strategy, wepropose to apply the Active Queue Management Algorithm—REM in FLM. By using thecongestion indication provided by REM, our purpose is to make the router detect thecongestion or spare state of network and report the state information to end systems. Based onthe received information, end systems adjust the synchronization points strategy for differentreceivers to coordinate their join and leave attempts. The information from routers is also used...