Incentives for Cooperation in Peer-to-Peer Networks
Incentives for Cooperation in Peer-to-Peer Networks
复制标题
DOI：
--
复制
发表时间：
2003
影响因子：
4.2
通讯作者：
Kevin Lai;M. Feldman;I. Stoica;J. Chuang
中科院分区：
文献类型：
--
作者：
Kevin Lai;M. Feldman;I. Stoica;J. Chuang
为了模拟这个问题，我们使用进化囚徒困境（EPD）[3]来捕捉个人和社会效用之间的紧张关系。在EPD所描述的情况下，合作需要重复和声誉。这两种技术都增加了实体之间的熟悉度（直接或间接），从而减少了与陌生人互动的可能性，从而加强了合作。然而，P2P应用的背景提出了新的挑战。首先，这些系统的大规模使得与同一实体发生重复交互的可能性较小。第二，信誉假设玩家保持持久的身份，但在许多P2P系统中零成本身份的存在允许实体不断改变身份。这两个问题都增加了与陌生人互动的可能性。
To model this problem, we use the Evolutionary Prisoner’s Dilemma (EPD) [3] to capture the tension between individual and social utility. In the situations that EPD characterizes, cooperation requires repetition and reputation. Both techniques increase familiarity between entities (either directly or indirectly), thereby reducing the probability of interactions with strangers and consequently enhancing cooperation. The context of P2P applications, however, imposes new challenges. First, the large scale of these systems makes it less likely that repeat interactions will occur with the same entity. Second, reputation assumes that players maintain persistent identities, but the existance of zero-cost identities in many P2P systems allows entities to continuously change identities. Both of these issues increase the probability of interacting with a stranger.