P2P流媒体直播业务流量优化研究

发布时间：2018-03-27 07:07

本文选题：点对点　切入点：流媒体直播　出处：《北京邮电大学》2014年硕士论文

【摘要】：P2P (Peer-to-Peer,点对点)技术是近些年来出现的一种可以高效解决网络中大规模内容分发问题的技术,采用P2P技术构建的P2P网络具有低成本和高度可扩展性的特点。在P2P网络中,用户节点同时扮演消费者和提供者的角色,用户节点通过分享自己的资源,构建了一个大型的用户资源池,来帮助实现一个只需要相对较小基础设施支持的全局性任务。 P2P流媒体直播业务作为基于动态性和实时性的P2P业务,受到人们越来越多的关注。为了达到流媒体直播业务的要求,必须对网络中的流量进行优化。现有一部分关于P2P流媒体直播业务的流量优化方案,都仍存在一些不足。如没有考虑节点的不稳定性,导致节点频繁的加入和离开会引起系统不稳定；流量本地化导致流量过于集中而增大网络拥塞的风险；数据调度过程中产生大量的冗余流量和开销等。针对现有方案中存在的不足,本文提出了一种P2P流媒体直播业务流量优化方案。在为系统中的节点选择邻居节点的过程中,将节点的在线时长添加到选择参数中,使得邻居节点选择更倾向于在线时间较长的节点。为了防止网络拥塞的出现,对节点和PID(Provider-Defined Network Location Identifier,提供商定义的网络区域标识符)的负载率进行监控,当节点的负载率达到阈值时便不再将它作为邻居节点进行分配,并周期性地检查PID的负载情况。为了减少数据调度过程中的冗余流量和开销,本文提出使用服务器对节点的缓冲区位图集中管理,记录每个节点中数据块请求情况的方法,将数据请求转发至最合适的目标节点。本文首先分析了现有P2P流媒体直播业务流量优化方案的特点及其缺陷,然后分几个章节描述了本文提出的邻居节点选择机制,拥塞控制机制和数据调度方案,最后给出了基于OMNeT++仿真环境对本文方案进行的实验结果和分析。
[Abstract]:P2P Peer-to-Peer (peer-to-peer) technology is a kind of technology that can efficiently solve the problem of large-scale content distribution in network in recent years. The P2P network constructed by P2P technology has the characteristics of low cost and high scalability. The user node plays the role of both consumer and provider. By sharing its own resources, the user node constructs a large pool of user resources to help realize a global task that only needs to be supported by relatively small infrastructure. As a P2P service based on dynamic and real-time, P2P streaming streaming service has attracted more and more attention. It is necessary to optimize the traffic in the network. There are still some shortcomings in some of the current traffic optimization schemes for P2P live streaming services, such as not considering the instability of the nodes, The frequent joining and leaving of nodes will lead to system instability; traffic localization will lead to excessive concentration of traffic and increase the risk of network congestion; data scheduling process will produce a large amount of redundant traffic and overhead. In view of the shortcomings of the existing schemes, this paper proposes a P2P streaming streaming traffic optimization scheme. In the process of selecting neighbor nodes for the nodes in the system, the online time of the nodes is added to the selection parameters. In order to prevent network congestion, load rates of nodes and PID(Provider-Defined Network Location identifiers, defined by providers, are monitored. When the load rate of the node reaches the threshold, it is no longer assigned as a neighbor node, and the load of PID is checked periodically. In this paper, a method of using server to manage the buffer bitmap of a node is proposed to record the data block request in each node and forward the data request to the most suitable target node. This paper first analyzes the characteristics and defects of the existing P2P streaming stream traffic optimization scheme, then describes the proposed neighbor node selection mechanism, congestion control mechanism and data scheduling scheme in several chapters. Finally, the experimental results and analysis of the scheme based on OMNeT simulation environment are given.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.02

【相似文献】