Research On Application Layer Multicast Robustness Based On P2P

Streaming media over the Internet is booming nowadays.IP multicast and application layer multicast has become the study hotspots.Because of the inherent limitations of IP multicast,it had not put in a large-scale commercial applications. Based on the P2P network of media distribution techniques,ALM has been extensively studied.The function of the multicast is transferred from network layer to the application layer.So the stability of the terminal determines the ALM,and the terminal failure or withdrawal of the multicast tree frequently,impacted on other users to receive multicast data's continuity seriously,weakened the robustness of the multicast. Therefore,research on ALM robustness based on P2P for improving the quality of service is important.At present,the scheme of enhancing robustness of the ALM include:reducing the frequency of failure nodes,failure detection and tree reconstruction technique,flood mechanism,multi-trees and rings.The goal of failure detection and tree reconstruction is to shorten the recovery time,is aslo an important research.Tree reconstruction includes proactive approach and reactive approach.Proactive approach has pre-constructed tree reconstruction scheme,can be relatively rapid recovery of multicast tree and enhancing robustness.This paper analyzes the various schemes,using failure detection and tree reconstruction technique,then we propose an algorithm of enhancing robustness of ALM, specifically as follows:1.Building a spanning tree of at least one residual degree based on P2P overlay network.It can be reconstructing after nodes failure when constructing the backup routes beforehand which based on the node residual degree in accordance with the minimum cost principle.When the nodes consume the degree,according to the node to brother node's response message,adjusting the position of the tree appropriately, guaranteed that the node has the redundancy.2.Combined with layered coding technique to improve the algorithm on the forwading plane:According to the streaming media transmission layer to limit the degree, adjusting the transmission media layers timely after degree consumption.Using the improved shared learning scheme to redefine the affected regions,it can reduce the unnecessary join-experiments of the other nodes.3.we evaluate the algorithm by experiments using the network simulator ns-2.The results indicate that our approach has shorter recovery time,lower control overheads, and the delivery delays is aslo under the reasonable control,so as to achieve the purpose of enhancing robustness of ALM.