基于合作博弈的P2P内容分发技术的设计和分析
发布时间：2018-12-16 17:36
【摘要】：P2P内容分发网络是融合了P2P网络和内容分发技术各自的优点而产生的，是目前互联网应用领域研究的热点之一。但是在实际的网络环境中存在着大量的自私节点，它们占用大量的带宽资源却又不愿意为其他节点提供下载服务，导致“搭便车”和“公共地悲剧”现象的产生。这违背了P2P最初的设计理念，严重影响了系统的整体性能。本文针对这一现象，引入了经济学中的合作博弈理论，利用合作博弈理论中的三种技术：Shapley值、NBS(Nash Bargaining Solution)和Repeated Game来解决P2P内容分发网络中激励节点合作实现资源共享的问题。 本文首先对P2P内容分发网络进行了综述，包括研究背景、研究现状和研究意义，重点讨论了目前已有的激励机制。接着介绍了P2P技术、内容分发网络(CDN)和博弈论等理论知识。最后针对目前激励机制在P2P内容分发网络中存在的问题，提出了基于合作博弈的内容分发技术的设计方案，目的在于抑制“搭便车”现象的发生和促进节点合作，从而提高P2P内容分发网络的整体性能。 本文基于合作博弈的P2P内容分发技术的设计和分析主要包括三个方面：首先，提出了基于Shapley值的P2P视频点播系统的存储管理机制。该机制通过采用Shapley值的方法实现了P2P视频点播系统中合作收益的公平合理分配，从而激励了节点参与合作，提高了随机搜索操作的命中率；其次，提出了基于NBS的社会P2P网络的内容分发管理方案。在社会P2P网络这种典型的P2P内容分发网络中，由于节点都具有社会属性，节点之间易形成合作联盟从而实现带宽共享，但是节点合作时的最优策略选择和节点的公平性等问题亟需研究。因此我们通过建立讨价还价博弈模型并求出了纳什谈判解，提高了节点的效率和公平性；最后，提出了社会P2P网络中基于重复博弈的惩罚激励机制。在社会P2P网络中，由于节点都是自私的，我们不能保证在整个合作过程中所有节点一直都是诚信的，针对怎样防止节点欺诈行为发生的问题，提出了惩罚激励机制“Punishment-Forgiving-Restoring(PFR)”。该机制有效地抑制了节点欺诈行为的发生，促进了节点的诚信合作并且提高了系统的平均收益。 仿真结果表明，基于合作博弈的P2P内容分发技术能够有效地抑制了“搭便车”行为，实现了P2P内容分发网络资源的合理分配，平衡了网络的效率和公平性，提高了网络的整体性能，，对推动P2P内容分发技术应用的进一步发展具有重要意义。
[Abstract]:P2P content distribution network, which combines the advantages of P2P network and content distribution technology, is one of the hotspots in the field of Internet application. However, there are a lot of selfish nodes in the actual network environment, they occupy a lot of bandwidth resources, but they are unwilling to provide download services for other nodes, which leads to the phenomenon of "hitchhiking" and "tragedy of common land". This violates the original design idea of P2P and seriously affects the overall performance of the system. Aiming at this phenomenon, this paper introduces the cooperative game theory in economics, and makes use of three techniques in the cooperative game theory: Shapley value, NBS (Nash Bargaining Solution) and Repeated Game are used to solve the problem of encouraging nodes to cooperate to share resources in P2P content distribution networks. Firstly, this paper summarizes P2P content distribution networks, including the research background, research status and significance, and focuses on the existing incentive mechanisms. Then it introduces P2P technology, content distribution network (CDN) and game theory. Finally, aiming at the existing problems of incentive mechanism in P2P content distribution network, this paper puts forward a design scheme of content distribution technology based on cooperative game, which aims at restraining the phenomenon of "hitchhiking" and promoting node cooperation. In order to improve the overall performance of P2P content distribution network. The design and analysis of P2P content distribution technology based on cooperative game includes three aspects: firstly, the storage management mechanism of P2P VOD system based on Shapley value is proposed. The mechanism realizes fair and reasonable distribution of cooperative income in P2P VOD system by using Shapley value method, thus motivating nodes to participate in cooperation and improving hit ratio of random search operation. Secondly, the content distribution management scheme of social P2P network based on NBS is proposed. In the typical P2P content distribution network of social P2P network, because the nodes all have the social attribute, it is easy to form the cooperation alliance among the nodes to realize the bandwidth sharing. However, the optimal policy selection and the fairness of nodes need to be studied. So we build a bargaining game model and find the Nash solution to improve the efficiency and fairness of the node. Finally we propose a penalty incentive mechanism based on repeated game in social P2P network. In the social P2P network, because the nodes are selfish, we can not guarantee that all the nodes in the whole cooperation process are always honest, aiming at how to prevent node fraud. The penalty incentive mechanism "Punishment-Forgiving-Restoring (PFR)" is put forward. This mechanism effectively inhibits the occurrence of node fraud, promotes the cooperative integrity of nodes and improves the average income of the system. The simulation results show that the P2P content distribution technology based on cooperative game can effectively restrain the "hitchhiking" behavior, realize the rational allocation of P2P content distribution network resources, and balance the efficiency and fairness of the network. It improves the overall performance of the network and plays an important role in promoting the application of P2P content distribution technology.
【学位授予单位】：苏州大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.02
本文编号：2382772
[Abstract]:P2P content distribution network, which combines the advantages of P2P network and content distribution technology, is one of the hotspots in the field of Internet application. However, there are a lot of selfish nodes in the actual network environment, they occupy a lot of bandwidth resources, but they are unwilling to provide download services for other nodes, which leads to the phenomenon of "hitchhiking" and "tragedy of common land". This violates the original design idea of P2P and seriously affects the overall performance of the system. Aiming at this phenomenon, this paper introduces the cooperative game theory in economics, and makes use of three techniques in the cooperative game theory: Shapley value, NBS (Nash Bargaining Solution) and Repeated Game are used to solve the problem of encouraging nodes to cooperate to share resources in P2P content distribution networks. Firstly, this paper summarizes P2P content distribution networks, including the research background, research status and significance, and focuses on the existing incentive mechanisms. Then it introduces P2P technology, content distribution network (CDN) and game theory. Finally, aiming at the existing problems of incentive mechanism in P2P content distribution network, this paper puts forward a design scheme of content distribution technology based on cooperative game, which aims at restraining the phenomenon of "hitchhiking" and promoting node cooperation. In order to improve the overall performance of P2P content distribution network. The design and analysis of P2P content distribution technology based on cooperative game includes three aspects: firstly, the storage management mechanism of P2P VOD system based on Shapley value is proposed. The mechanism realizes fair and reasonable distribution of cooperative income in P2P VOD system by using Shapley value method, thus motivating nodes to participate in cooperation and improving hit ratio of random search operation. Secondly, the content distribution management scheme of social P2P network based on NBS is proposed. In the typical P2P content distribution network of social P2P network, because the nodes all have the social attribute, it is easy to form the cooperation alliance among the nodes to realize the bandwidth sharing. However, the optimal policy selection and the fairness of nodes need to be studied. So we build a bargaining game model and find the Nash solution to improve the efficiency and fairness of the node. Finally we propose a penalty incentive mechanism based on repeated game in social P2P network. In the social P2P network, because the nodes are selfish, we can not guarantee that all the nodes in the whole cooperation process are always honest, aiming at how to prevent node fraud. The penalty incentive mechanism "Punishment-Forgiving-Restoring (PFR)" is put forward. This mechanism effectively inhibits the occurrence of node fraud, promotes the cooperative integrity of nodes and improves the average income of the system. The simulation results show that the P2P content distribution technology based on cooperative game can effectively restrain the "hitchhiking" behavior, realize the rational allocation of P2P content distribution network resources, and balance the efficiency and fairness of the network. It improves the overall performance of the network and plays an important role in promoting the application of P2P content distribution technology.
【学位授予单位】：苏州大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.02
【参考文献】
相关期刊论文 前6条
1 刘业;杨鹏;;基于重复博弈的P2P网络信用管理机制的研究[J];计算机研究与发展;2006年04期
2 乐光学;李仁发;陈志;周旭;;P2P网络中搭便车行为分析与抑制机制建模[J];计算机研究与发展;2011年03期
3 刘浩;卿桐;张连明;;一种P2P网络的重复博弈激励机制[J];计算机工程与应用;2010年35期
4 刘瑜;朱光喜;尹浩;刘雪宁;;内容分发网发展综述[J];计算机科学;2009年02期
5 余一娇;金海;;对等网络中的搭便车行为分析与抑制机制综述[J];计算机学报;2008年01期
6 徐海湄;郑相全;齐守青;聂晓文;;一种新的基于博弈论的P2P激励机制[J];计算机应用研究;2008年09期
本文编号：2382772
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2382772.html
