摘要
安全多方计算为保护各方的私有信息,采用安全协议来保证合作计算的顺利进行。但恶意攻击的存在,使得安全协议的复杂性较高,协议的可操作性较低。鉴于此,提出一种基于声誉的评分累积信任模型,根据参与节点的历史行为评估其声誉,辨别恶意节点,采用惩罚机制鼓励可信的参与节点、隔离不可信节点,从而降低恶意攻击带来的风险。实验表明,该模型可以在一定程度上抵制自私的恶意攻击。
In order to preserve privacy, secure protocols are adopted to do multi-party computation. They become much more complex because of kinds of malicious attacks and have lower operability. In consideration of these facts, an accumulated score trust model based on reputation is presented, which can discover malicious nodes ac- cording to the reputation from nodes' history behaviors. A punishment mechanism is also adopted to encourage the trust nodes and isolate distrust nodes, which reduce the risk from attacks. The experimental results show that this model can prevent selfish malicious attack in a certain degree.
出处
《计算机工程与应用》
CSCD
2012年第27期69-73,共5页
Computer Engineering and Applications
基金
国家自然科学基金(No.60703071)
安徽高校省级自然科学研究重点项目(No.KJ2010A133)
安徽省高等学校青年人才基金项目(No.2011SQRL026)
安徽省高校省级科学研究项目(No.KJ2011Z142)
关键词
安全多方计算
恶意
声誉
自私
惩罚机制
secure multi-party computation
malicious
reputation
selfish behavior
punishment mechanism