Abstract: Incentive mechanism plays a significant part in BitTorrent protocol. It encourages peers to upload more resources in order to obtain better download rate, and improves the whole system sharing performance. However, the original incentive mechanism in BitTorrent protocol allows such kinds of free-riding nodes, which bring bad incentive effects. This paper proposes a novel incentive mechanism with detailed algorithm for its practical application. In order to make more accurate and sound analysis for this novel incentive mechanism, this paper firstly constructs its colored Petri nets (CPN) based formal model, then validates the accuracy of this model, finally carries out the performance analysis for the BitTorrent protocol with new proposed incentive mechanism using model simulation. The performance analysis results show that the novel incentive mechanism can provide more strict restriction to free-riding node behaviors and archive better incentive effects.

Key words: BitTorrent, incentive mechanism, colored Petri nets (CPN), model simulation

摘要： 激励机制是BitTorrent协议的重要组成部分，它鼓励用户上传以获得更快的下载速度，并提高系统整体性能。原有BitTorrent激励机制存在大量的搭便车节点，极大影响了激励效果，因此提出了一种新的激励策略，并给出了具体的实施算法。对应用这种新激励机制的BitTorrent协议进行了详细的着色Petri网（colored Petri nets，CPN）形式化建模分析，在验证模型正确性的基础上，进行基于模型模拟的性能分析实验，结果表明该激励机制能够严格限制搭便车节点行为，达到更好的激励效果。

关键词: BitTorrent, 激励策略, 着色Petri网（CPN）, 模型模拟