An Incentive Mechanism for P2P Networks Based on Knowledge Diversity

• Main Authors: Yu,Su-Chin, 余素錦
• Other Authors: Gen-Yih Liao
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/56277960009836083452

碩士 === 長庚大學 === 資訊管理研究所 === 95 === The Peer-to-Peer (P2P) networks offer data exchange model that is different from traditional Client-Server architecture. Since there is no central control servers in the P2P networks, P2P has provided a more convenient way for data exchange and retrieving. If most users lack the motive of sharing resource in existing P2P networks, will cause P2P networks to move towards becoming feeble and die. As a way to allow the P2P network to operate permanently, incentive mechanism has recently been introduced into the P2P system. It attempts to attract more system users and more resource providers to enrich the network. In existing papers, the knowledge sharing based of P2P system that shared the hard disk space and bandwidth, the design of its incentive mechanism is designing with the bandwidth or file number. This thesis thinking, diversity of resources more important than that knowledge sharing based of P2P system, the diversity of resources is more important than the sharing of hard disk space and bandwidth. Besides, this research still assumes whether network members of P2P have a close interest, will effect the management of P2P network. Therefore, in order to verify the importance of resource diversity and interest clustering degree, we study relevant papers, put forward relevant research and assume, utilize the simulation experiment way to examine finally. The experiment result has proven the proposed research hypothesis: the incentive mechanism of resource diversity has a greater significant impact to user effect. On the interest clustering degree aspect, as using supply satisfaction to measure, we assume that interest clustering high degree is relatively significant to user effect；as using demand utility to measure, we assume that interest clustering low degree has a greater significant effect to user.