摘要
为解决P2P网络中节点的不合作行为和恶意攻击等问题,提出了一种分布式兴趣信任模型DITM.模型通过划分兴趣域聚集兴趣相似的节点来解决节点间因兴趣不对称难以建立直接信任关系的问题.利用兴趣相似度刻画节点在其偏好领域上的服务行为相似性,并通过兴趣相似度加权域推荐信任度计算域服务信誉.节点在不同域内的服务信誉组成一个服务信誉向量,该向量维护了节点在各个兴趣域的服务行为细节,从而能有效抵御恶意节点针对特定兴趣域的攻击,并能激励好节点在多个域内贡献资源.仿真实验表明,DITM较已有的信任模型在迭代的收敛速度、下载成功率和模型的安全性等问题上有较大提高.
To solve the uncooperative behavior and malicious attack problems of nodes in P2P(peer-to-peer) networks,a distributed interest trust model(DITM) is proposed,which clusters nodes with similar interest into the same interest domains,so that the hardness of constructing direct trust relationship between nodes due to asymmetric interest feature is solved.The DITM uses interest similarity to describe nodes service similarity in their biased domains,and uses it to weight recommendations of different nodes when calculating domain service trust of a specific node.Service trust values of a node in different domains form a service trust vector,which describes the service details of the node in each domain.By this means,the DITM can defend attacks of malicious nodes on specific domain,and provide incentives for good nodes to contribute resources in each interest domain.Simulation results show that the convergence time,download success ratio and security of the DITM are remarkably improved compared with the current typical trust model.
出处
《东南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2011年第2期242-246,共5页
Journal of Southeast University:Natural Science Edition
基金
江苏省自然科学基金资助项目(BK2010133)
关键词
对等网络
兴趣域
域推荐信任度
域服务信誉
兴趣相似度
peer to peer network
interest domain
domain recommendation trust
domain service reputation
interest similarity