BitTorrent激励机制的建模与分析  

Modeling and analysis of BitTorrent incentive mechanism

作　　者：李晓义[1] 李治军[1] 姜守旭[1] 

机构地区：[1]哈尔滨工业大学计算机科学与技术学院,哈尔滨150001

出　　处：《哈尔滨工业大学学报》2011年第11期65-69,共5页Journal of Harbin Institute of Technology

基　　金：国家自然科学基金资助项目(60803148)

摘　　要：为了考察影响BitTorrent激励机制的因素,进行了建模和分析.针对激励机制存在随机选择这个特点,采用了马尔可夫过程进行建模.建模中首先得出转移概率矩阵,然后计算转移概率矩阵的极限.理论分析及实验表明:相同上传带宽的节点会聚类,通过调整随机选择的节点可以影响聚类速度.该模型符合激励机制的运行过程,可以通过改变模型中的相应参数来改善系统性能.To investigate the factors that affect incentive mechanism of BitTorrent system, this paper makes a modeling analysis. Because there are random selections, we use Markov process to model the incentive mechanism. We first get the transition probability matrix and then calculate the limit of the transition probability ma- trix. Theoretical analysis and experiment results show that peers with same upload bandwidth cluster, and ad- justing randomly selected nodes can affect the speed of clustering. The model is consistent with the incentive mechanism, and by adjusting parameters we can improve the performance of the system.

关 键 词：BITTORRENT系统 激励机制 马尔可夫过程 聚类 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]