摘要
现有的P2P系统信任评价模型正面临着两种恶意节点的攻击行为——策略性欺骗和不诚实推荐,严重影响了模型计算节点信任评价的准确性和有效性.针对现有模型存在的不足,提出了一种基于概率统计方法的信任评价模型.该模型借鉴人类社会中主观信任关系的概念,依据直接经验和反馈信息,利用概率统计方法分别计算节点的直接信任和推荐信任,并通过区分直接经验的重要程度,区分反馈信息及其推荐者的可信度,提高信任评价模型的有效性.仿真实验分析说明,与已有的信任评价模型相比,该模型能够更有效地抑制策略性欺骗和不诚实推荐的威胁,特别是复杂的协同作弊方式对系统的攻击.
Peer to peer （P2P） technology has been widely used in file-sharing, distributed computing, emarket and information management. One of the fundamental challenges for P2P systems is the ability to manage risks involved in interacting and collaborating with prior unknown and potentially malicious parties. Reputation-based trust management systems can successfully mitigate this risk by deriving the trustworthiness of a certain peer from that peer＇s behavior history. However, in current trust models employed by the existing P2P systems, the validity of peers＇ trust valuation is seriously affected by peers＇ malicious behaviors. For example, there are many strategic cheating and dishonest recommendation. To solve this problem, a novel reputation-based trust model based on probability and statistics for P2P systems is proposed. Referring to subjective trust relationship of sociological theory, the proposed model uses experience-based and recommendation-based trust relationship to compute the trustworthiness of peers. In particular, this model introduces three parameters, namely, experience＇s time-sensitivity, referee＇s credibility and recommended information＇s reliability, and thus it can provide adequate reaction to peers＇ malicious behaviors. Theoretical analysis and simulation show that the proposed model has advantages in coping with peers＇ malicious behaviors over the existing P2P reputation systems. It is highly effective in countering malicious peers regarding strategic cheating, dishonest recommendation, and collusion.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2008年第3期408-416,共9页
Journal of Computer Research and Development
基金
国家“八六三”高技术研究发展计划基金项目(2007AA01Z422)
国家发改委高技术研究发展计划基金项目(CNGI-04-16-18)
关键词
对等网络
推荐
信任
恶意行为
协同作弊
peer to peer
recommendation
trust
malicious behavior
collusion
