P2p Coverage Multicast Network Study

As the rapid development of computer and network technology, the ability of transmitting and processing of Internet has been improved greatly, many new applications with multimedia features have emerged in network. For example, video-conferencing, distance education, network multicast and so on. These multimedia applications have higher latency, larger data, and longer time. Moreover, they require higher delivery bandwidth and stronger real-time property. If these problems are solved by broadcast and repeated unicast, it will waste network bandwidth seriously, so many researchers began to study multicast.Multicast is a technology of transmitting a data packet from one sender (also called multicast source) or many senders to many receivers. Multicast can make one or many multicast sources transmit data packets to the special multicast group, only the hosts who have joined this special multicast group can receive the data packets. Multicast saves network bandwidth, reduces network loads and avoids "broadcasting-storm". Multicast can be divided into two categories called IP multicast and application layer multicast. IP multicast has a lot of disadvantages, so it can not be widely used in Internet. But many new applications with multimedia features require that Internet can support multicast eagerly, so many researchers proposed multicast.In the same time, P2P technology develops greatly, and P2P overlay networks have a lot of advantages, so application layer multicast based on P2P overlay networks becomes the hot issue recently. The classic application layer multicast based on P2P overlay networks can be listed as follows:multicast based on Chord, CAN-multicast, Scribe and Bayeux.We study multicast in P2P overlay networks in this thesis. This thesis can be divided into six chapters. The first chapter describes study background and significance, research status of multicast, study target and content of this thesis. In addition, we arrange the organizational structure of this thesis in first chapter. The second chapter describes P2P overlay network and application layer multicast. The third chapter describes the enhanced NEMO protocol. We enhanced NEMO which is a performance-centric, overlay layer multicast protocol in the aspect of cluster leaders selecting and data forwarding. The simulation results show that the enhanced NEMO protocol is better than NEMO protocol in the average delivery latency and accumulative failure recovery overhead. The fourth chapter describes DPM:a magnetic disk-based P2P multicast protocol. We illustrate the construction of DPM hierarchy, the processes of node joining, node leaving, node maintenance and DPM multicast. We make a contrast between DPM and Chord multicast by simulation. The results show that multicasting on DPM has less delivery latency than that of on Chord under the same network size. What is more important, DPM has better scalability and can accommodate more nodes. It can be used in large-scale multicast applications. The fifth chapter describes CONE-M:an application layer multicast model based on cone. We illustrate the construction of Cone-M, the processes of node joining, node leaving and Cone-M multicast. We analyze Cone-M by simulation. The results show that the average link stress on Cone-M is lower than that of on NICE, and the average data delivery ratio on Cone-M is higher than that of on NICE in the process of multicasting. What is more important, Cone-M avoids free-riding which exists in NICE by making use of every node in system. The sixth chapter summarizes this thesis and prospects the future work.