一种双重度量的分布式P2P信任模型

Distributed P2P trust model based double metric

现有的P2P信任模型通常面向节点行为计算其信誉值,不能处理高信誉节点非法上传的特殊情况。提出一种双重度量信任模型,引入聚类方法分派节点的推荐权重,通过迭代得到节点的全局信誉值,然后结合资源自身的信誉信息,给出节点选择下载源的方法。仿真分析表明,相比类似模型,该模型可以进一步缓解节点恶意行为带来的负面影响。

In existing P2P trust models,the reputation values are always computed for behavior of peers.They can not deal with such particular case as that peers with high reputation value upload inauthentic resources.A new trust model is presented with double metric.The cluster method is introduced for designing the recommendation weight.The global reputation values of peers are gained by iteration.Associated with the reputation of peers and the credibility of resources,the methods of selecting peers for download are given.Simulation analyses show that,this model can alleviate the negative effect which is caused by peer＇s malicious behavior more effective than parallel models.