基于关系分解的外包数据库机密性保护方案研究

Relation-Fragmentation-Based Confidential Scheme for Database Outsourcing

针对外包数据库服务中数据库服务提供商本身并不完全可信等潜在威胁引起的隐私数据机密性威胁,提出了一种基于关系分解的外包关系数据库机密性保护方案。设计了考虑查询类型和属性所处查询位置的属性分解代价计算策略,并以此计算策略为评价标准,给出了求解最小属性分解代价方案的算法。然后,对各数据切片进行了准标识符的抗共谋攻击处理,并进行了性能测试。实验对比结果表明该方案对不同的应用查询类型有较好的适应性,有效降低了关系分解代价,提高了查询执行效率。

A confidentiality protecting scheme based on relation fragmentation is proposed to solve the privacy risk caused by untrusted database service providers （DSP） in database outsourcing sce- nario. The scheme firstly designs an application oriented strategy to compute attribute fragmentation cost, which concerns the query type and the position that the involved attribute is in. Secondly, by using the self-adaptive strategy, an algorithm is established to find the least cost attribute fragmenta- tion scheme for the given table. Furthermore, the link of identifiers is broken of the same tuples in different fragmented attribute sub-set to resist DSPs-collusion. Finally, a brief look is given of the query executing mechanism for this scheme and an experiment is demonstrated to test its efficiency.