Font Size: a A A

Research On Nodes Churn Of Streaming Media Living Broadcast System

Posted on:2011-11-26Degree:MasterType:Thesis
Country:ChinaCandidate:J F YuanFull Text:PDF
GTID:2178330332958943Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
In recent years,with the popularization and development of the network,a large number of P2P streaming media applications have begun to emerge.P2P live streaming technology as an important branch of streaming media has become a hot topic of research.Based on PeerCast systems,the project team developed LStream which is a P2P system of a large-scale live video.LStream system uses a hierarchical structure.In the node scheduling module project team member propose a web-based coordinates of GNP combining with Pcoord node scheduling strategy. In the transmission and the data scheduling module the system uses an incentive-based push-pull strategy of combining data and LRFA algorithm.The focus of this article is how to design a new anomaly detection mechanisms to ease the current disturbance issues of P2P live streaming system.In the current P2P video broadcast systems there are four anomaly detection mechanisms.They are central detection mechanism,strong node detection mechanism,father and son nodes detection mechanism and supervision mechanism.However,several mechanisms are all used conventional heart ways.They have some drawbacks.For example anomaly detection cycle is too long,the node can't detect regional failure and system redundancy controlling packets are too much and the controlling load rate is too high.Combining with the features of different layers nodes of LStream system.The paper designs an abnormal detection mechanism.This mechanism is composed of a variety of mechanisms.It includes cache data detection (Buffer Data),remote monitoring and BM(Buffer Map) heartbeat detection mechanism.The total detection mechanism is known as BD-BM mechanism.Because SP(Super Peers) nodes have the only source of data,so the system uses cache data detection mechanism.To detect the whole failure of SNP (Super Normal Peers) nodes,the system increase the remote monitoring for the SNP nodes.As NP nodes use PULL data transfer mechanism.Neighbor nodes need to send each other BM information (this information is to tell neighbor node that the node has the information of data pieces).So it use the BM heartbeat mechanism.The BD-BM detection mechanism can shorten the time of anomaly detection nodes and detect regional failure of the nodes. It can also reduce control packets of the system and reduces controlling load rate of the system.
Keywords/Search Tags:P2P, livebroadcast, streaming media, churn, Anomaly detection
PDF Full Text Request
Related items