大数据环境下的多维敏感度最佳k值匿名框架  

MULTI-DIMENSIONAL SENSITIVITY BEST K-VALUE ANONYMITY FRAMEWORK IN BIG DATA ENVIRONMENT

作　　者：南楠 严英占[2] Nan Nan;Yan Yingzhan(College of Basic Education,Lingnan Normal University,Zhanjiang 524048,Guangdong,China;The 54th Research Institute,China Electronics Technology Group,Shijiazhuang 050000,Hebei,China)

机构地区：[1]岭南师范学院基础教育学院,广东湛江524048 [2]中国电子科技集团第54研究所,河北石家庄050000

出　　处：《计算机应用与软件》2020年第6期297-302,共6页Computer Applications and Software

基　　金：河北省教育厅青年基金项目(QN2016182)。

摘　　要：针对现有匿名算法未提供多维敏感度和细粒度的隐私保护问题,提出大数据的多维敏感度的最佳k值匿名(Multi-dimensional sensitivity optimal k-value anonymous,MSOkA)框架。通过敏感度的计算对用户访问权限分级,对准标识符QI敏感数据聚合分组,提供多维敏感度的细粒度用户访问,根据不同的授权级别进行权限调整。根据中断方程,设计累积频率(cumulative frequency,CF)得到最佳k值,通过选择最佳所有权级别来确定访问粒度,减少明显猜测和跨组唯一标识符攻击。另外,设计一个在MapReduce环境中运行的隐私框架,在hadoop域上实现框架的核心服务、初始化服务和匿名服务,通过SAML连接联合身份验证服务和服务提供者,实现对大数据的多维快速隐私保护。实验结果表明,该方法能够找到最佳k值,实现大数据多维敏感度和细粒度的隐私保护,且隐私保护程度和执行时间都优于现有方法。Aiming at the problem that the existing anonymity algorithms do not provide multi-dimensional sensitivity and fine-grained privacy protection,this paper proposes the multi-dimensional sensitivity optimal k-value anonymous(MSOkA)framework for big data.Through the calculation of sensitivity to the user access level,the sensitive data of the identifier QI was aggregated and grouped to provide fine-grained user access with multi-dimensional sensitivity,and the permissions were adjusted according to different authorization levels.According to the interrupt equation,the cumulative frequency(CF)was designed to get the best k-value.And the access granularity was determined by selecting the optimal ownership level to reduce obvious guessing and cross-group unique identifiers attacks.In addition,a privacy framework was designed to run in the MapReduce environment,which implemented the core services,initialization services and anonymous services of the framework on the hadoop domain.The SAML connection was used to connect the federation service and the service provider to realize the multi-dimensional and fast privacy protection of big data.The experimental results show that the proposed method can find the best k-value and realize the multi-dimensional sensitivity and fine-grained privacy protection of big data,and the degree of privacy protection and execution time are better than other existing methods.

关 键 词：最佳k值 多维敏感度 细粒度 累积频率 MAPREDUCE 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]