Research On P2P Based Live Streaming Media Dissemination

While multimedia streaming over the Internet is booming nowadays, it continually gives new challenges to the distribution of streaming contents. The traditional internet model based on the client-server framework can't support on the large scale of the concurrent distribution because of the bottleneck at the server. The CDN network and IP multicast also can't be universalizable due to the expensive fare and technological problem.In order to improve the QoS of Streaming over the internet, media service over P2P network is proposed. It implement the mutlicast on the application level and carries the dissemination from streaming server to the users' network,which makes the streaming media service more expandable,more robust and more practical.Recently, the proctol of data dissemination which is applicable to the streaming media service mainly refers to two types: Tree-based protocol and Gossip-based protocol. Reserchers has proposed many related work and theoretical model.This paper presents a new P2PSTREAMING model: Coolice. It tries to achieve a better trade-off among bandwith efficiency ,delay and reliability by leveraging the advantages in both P2P and C/S. Coolice distribute the media data on the unstructured P2P network ,and manage all nodes through the concentrated server, which composes a simple and pratical solution for Gossip. Coolice has several salient features as follows: l)it make a good solution for the peer management by means of the advantage of C/S;2)it makes the information transition between peers efficient and simple,the P2P network of Coolice keeps in a unstructured topology with high throughput ,load balance and robustness;3)the architecture of the data dissemination system is clear and precise ,as the media source is separated from the server to be a particular peer participating in the dissemination;4)the packet scheduling algorithm based on the slidding window reduces the redundancy of the network and bring down the delay of packets, so that improve the QoS quality of the whole system.