跨关系云数据库属性名与密文等值连接共享协议研究

发布时间：2018-03-28 06:27

本文选题：云数据库　切入点：信息共享　出处：《西北农林科技大学》2017年硕士论文

【摘要】：云数据库是指部署和虚拟化在云计算环境中的数据库,通常由云服务提供商(Service Provider,SP)提供给不同用户。由于SP可能会在维护和管理数据库的过程中泄露用户隐私信息,用户需要将隐私数据加密后存放到云数据库中。虽然加密数据可以有效保护隐私数据,但是给用户间的信息共享造成了困难。现有的云数据库隐私信息共享方案存在没有属性名共享协议和等值连接共享协议效率太低等问题。在半诚实模型下,论文研究跨关系云数据库(Cloud Relational Database,CRDB)中两个用户间的加密信息共享。具体研究内容包括:(1)结合云数据库的特点,分析在保护隐私信息条件下共享面临的安全威胁,研究了针对云数据库用户的加密信息共享方案;详细介绍了方案的参与者、架构和执行流程;方案的执行流程包括用户连接、数据库连接和执行共享协议等三大部分。(2)提出了基于局部敏感哈希(Locality-Sensitive Hashing,LSH)的属性名共享方法;使用simhash算法实现了一个AS-sim协议。实验分析表明,当用户提供的数据库属性名长度大于9,设定阈值在60%到70%之间,属性名共享的准确率和召回率均可达到90%以上。(3)基于可交换加密函数提出了一种高效等值连接共享的EAVS协议,使用可交换加密函数计算某属性上的交集,共享信息计算双方通过共有属性值和对应元组的笛卡儿积,实现等值连接共享。安全分析表明,在半诚实模型下可实现2个用户间公平共享等值连接;效率分析表明,可显著提高等值连接共享的效率。
[Abstract]:A cloud database is a database that is deployed and virtualized in a cloud computing environment, usually provided to different users by a cloud service provider, Service provider SP.Because SP may disclose user privacy information while maintaining and managing the database, The user needs to encrypt the privacy data and store it in the cloud database. Although the encrypted data can effectively protect the privacy data, However, it is difficult to share information among users. The existing cloud database privacy information sharing schemes have many problems, such as the lack of attribute name sharing protocol and the low efficiency of equivalent connection sharing protocol. This paper studies the sharing of encrypted information between two users in a cross-relational cloud database, Cloud Relational Database (CRDB). The specific research contents include: 1) combined with the characteristics of cloud database, the paper analyzes the security threats faced by sharing under the condition of protecting privacy information. The encryption information sharing scheme for cloud database users is studied. The participants, architecture and execution flow of the scheme are introduced in detail. The execution process of the scheme includes user connection. This paper proposes a property name sharing method based on locally sensitive hashing Locality-Sensitive LSHs, and implements a AS-sim protocol using simhash algorithm. When the length of the database attribute name provided by the user is greater than 9, and the threshold is set between 60% and 70%, the accuracy and recall rate of attribute name sharing can reach more than 90%.) based on the exchangeable encryption function, an efficient and equivalent connection sharing EAVS protocol is proposed. Using the exchangeable encryption function to calculate the intersection on a certain attribute, the shared information is calculated through the Cartesian product of the shared attribute value and the corresponding tuple, and the equivalent connection and sharing are realized. The security analysis shows that, Under the semi-honest model, the fair sharing equivalent connection between two users can be realized, and the efficiency analysis shows that the efficiency of equivalent connection sharing can be improved significantly.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13;TP309

【相似文献】