一种基于服务期望的P2P动态信任模型  

A P2P Dynamic Trust Model Based on Service Expectation

作　　者：朱杰[1,2] 曹晓梅[1,2] 郭华娟[1,2] 

机构地区：[1]南京邮电大学计算机学院,南京210003 [2]南京邮电大学江苏省无线传感网高技术研究重点实验室,南京210003

出　　处：《计算机工程》2016年第3期177-181,共5页Computer Engineering

基　　金：国家自然科学青年基金资助项目"移动计算环境下新型僵尸网络分析与检测方法研究"(61202353)

摘　　要：为解决现有信任模型在节点行为改变时动态适应能力不足的问题,以及更好地应对网络中多样化的恶意攻击行为,提出一种P2P动态信任模型。引入能力评价、相似度、反馈可信度、服务期望、近期信任和历史信任等参数计算节点信任度,并借鉴计算机网络中拥塞控制思想缓慢上升、快速下降的原则,在反馈可信度的计算过程中加入奖惩机制,通过惩罚因子和时间衰减函数动态调节上述参数,提高信任模型的动态适应和抗攻击能力。实验结果表明,该模型能够真实地模拟现实信任关系,具有更好的动态适应能力,并可有效地抵御恶意节点威胁。In order to solve the poor dynamic adaptability to the node＇s behaviors in current P2 P trust models and deal with the variety of malicious attacks in the network,a newP2 P dynamic trust model is proposed. Several parameters are introduced in computing trust worthiness of peers,including capacity evaluation,similarity,feedback credibility,service expectation,recent trust,historical trust and so on. Together,referring to the idea of congestion control in computer network,additive increase / multiplicative decrease,a mechanism of rewards and punishments is introduced in the calculation of feedback credibility. These parameters are adjusted in time to reflect the dynamics of the trust environment using time attenuation factor and time decay function. The trust evaluation is able to simulate the real trust relationship and has better adaptability to the dynamics of trust. Theoretical analysis and simulation showthat the model has advantages in modeling dynamic trust relationship. It is highly effective in suppression of malicious attacks.

关 键 词：信任模型 服务期望 惩罚因子 时间衰减函数 近期信任 历史信任 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]