DRT:一种云计算中可信软件服务的通用动态演变鲁棒信任模型  

DRT:a Dynamical and Robust Trust Model for Software Service in Cloud Computing

作　　者：任伟[1] 雷敏[2,3] 杨榆[2] 

机构地区：[1]中国地质大学(武汉)计算机学院,武汉430074 [2]北京邮电大学网络与信息攻防技术教育部重点实验室,北京100876 [3]北京邮电大学软件学院,北京100876

出　　处：《小型微型计算机系统》2012年第4期679-683,共5页Journal of Chinese Computer Systems

基　　金：中央高校基本科研业务费专项资金项目(090109)资助;国家"九七三"重点基础研究发展计划项目(2007CB311203)资助;国家自然科学基金项目(60821001)资助;高等学校博士学科点专项科研基金项目(20070013007)资助

摘　　要：提出一种云计算的软件服务的信誉度量方法-动态演变的鲁棒信任模型DRT.DRT模型的优势在于:能有效地刻画动态特征,包括加权评级变化,时间演变,记忆平滑处理;信誉评级方法合理,包括依信誉权重的直接评级和间接评级;健壮性高,能抵御Bad-mouth、Id-shift、Collusion、Trust-abuse及其联合攻击.并提出DTAA攻击检测算法来检测Collusion攻击(以及联合攻击).模拟实验表明,关键信誉评级方法高效可行.A model for evaluating the repudiation of software service in cloud computing - a dynamical robust trust model DRT, is proposed. It can describe the dynamics of reputation including weighted grading alternation, time evolution, and memory smoothing process. It also consists a reasonable grading system, including weighted directly grading and indirectly grading. Moreover, the ro- bustness of the model performs well in that it can defend Bad-mouth, Id-shift, Collusion and Trust-abuse attacks { and their combina- tions ）. A detection algorithm - DAAT is also proposed for detection of Collusion attack （ and combination attack }. The simulation justified that the grading system is efficient and feasible.

关 键 词：可信软件 信任模型 云计算 软件服务 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]