|
With the increasing demand of the users of multimedia content , the traditional C / S structure streaming service can not meet a large number of users online at the same time , and not easy to form a system bottleneck , scalability . IP multicast technology inherent limitations , it is difficult to widely deploy . CDN distribution network costs are too high and only the bottleneck , the application surface is smaller . P2P network node receiving services to provide services to take full advantage of the free resources on the network , so service decentralized so as to reduce the load on the server , to improve the scalability of the system , to achieve large-scale streaming media data distribution hope. P2P streaming is mainly focused on the live system , less demand system . This paper analyzes the characteristics of P2P streaming technology node selection mechanism , summarized the key technologies of data transmission strategies as well as incentives . After analysis and comparison of the typical model of P2P streaming media , the results show that , with higher reliability and greater efficiency than the tree model based on the the Gossip protocol of P2P streaming model . Based on the above , we design a P2P video -on-demand prototype system based on Gossip protocol , custom system in Gossip interactive message types and formats , and the system 's caching mechanism and scheduling mechanism design , the use of fixed cache combined with dynamic cache cache management strategies to reduce the initial startup delay of the system , while scheduling strategy based on data-driven , and some improvements to ensure the timely arrival of data . The article also detailed description of the nodes in the system management strategies , proposed the concept of \Finally, compare the existing P2P simulator , select NS2 based simulation platform to further build a simulation platform for P2P simulation , simulation system based on program design , implementation, and performance testing . The test results show that the system is feasible and has some excellent scalability, stability, and efficiency .
|