Application Layer Multicast Forwarding Tree Node Selection Strategy

With the development of Internet, the one-to-many media streaming data distribution in large-scale is attracting more and more Internet users. The hotspot is the research of application layer multicast in live media streaming. Maintaining the nodes is one important issue in most ALM protocols, and the key process in that is how to select its parent node. The purpose of this thesis is to research how to select the optimal parent node when joining the multicast tree and bring forward a new node selecting algorithm, finally make a comparison with other two selecting algorithms.This thesis introduces some different techniques of selecting optimal parent node in most ALM protocols, and makes a deep analysis on that. In order to build a high efficient multicast tree, the research puts forward the distributed spanning tree algorithm using the Nearest network topology node selection strategy, which is based on a new group communication model named MixCast. The average delay and average joining time of the tree are analyzed. Simulation experiments show that this strategy takes more physical network circumstances into consideration, so it has high data communication efficiency.Different node selecting algorithm has different effect on the data communication efficiency, so the research is focused upon the influence on the communication efficiency of the multicast tree using three kinds of node selection strategy, which are based on Min-delay, Max-bandwidth, and Nearest network topology separately. The degree of the nodes is limited to build the multicast tree at different network scale. Simulation experiments show that the strategy based on Min-delay is the lowest in average delay and average joining time, and is much more suitable for the delay-sensitive application.