摘要
移动P2P网络的开放性和松耦合特性使得节点恶意攻击行为普遍存在,而现有基于声誉的信任模型大都基于"信誉值高的节点评价推荐越可信"的假设,无法识别恶意节点动态策略性攻击行为。针对这一问题,将社会网络相关理论引入信任系统,提出一种基于社会距离的信任模型(SD2Trust)。该模型区分了服务可信度和推荐评价可信度,用多维结构同型性描述向量刻画节点网络地位和行为特征,根据社会距离确定推荐节点集和推荐信誉计算权重,综合信任考虑了诋毁风险。理论分析和实验结果表明,该模型能有效对抗恶意节点动态策略攻击行为。
Most existing reputation-based trust models are based on the controversial assumption that ＂the higher serving credit one has, the more credible comments it provides＂. This paper proposes a Social-Distance based Distributed Trust model （SD2Trust） for mobile P2P networks, which differentiates the reliability of recommending and that of service providing, constructs trustworthy node set and computes reputation computing weights by making use of social similarity. The distributed reputation storing struc- ture and additional incentive approach together make it more suitable to mobile environment. Both theoretical analysis and simu- lation results show that SD2Trust can resist dynamic attacks with strategy and improve the success rate of transactions.
出处
《计算机工程与应用》
CSCD
2013年第23期75-80,共6页
Computer Engineering and Applications
基金
国家863主题项目(No.2011AA010604)
国家科技重大专项课题(No.2009ZX03004-002)
