基于最佳桶划分策略的密文索引技术

Cryptograph Index Technology Based on Strategy of Optimal Bucket Partitioning

在过去的几十年中,人们一直为寻求适用于数据库系统且安全高效的数据加密方法而努力,取得了一些研究成果,但到目前为止仍然没有解决加密后秘密查询命中率低、后处理代价大和密文索引安全性差的问题.本文将视图安全的判定条件k-匿名引入到了对密文索引列安全性的判定中,并针对由于构建密文索引而造成的信息泄露问题,提出了通过构造哈希函数和对多敏感属性统一进行桶划分并建立密文索引两种解决方案.最后提出了一种最佳桶划分策略,在桶划分子区间数目根据安全要求被限定的情况下,利用该桶划分算法建立密文索引可获得最优的查询命中率,从而平衡了密文索引的安全性和查询效率的问题.

People always made efforts on pursuing high-efficiency encryption method to adapt to database system during past decades. Although achievements have been made, there still are problems of long responding, expensive cost and low security of cryptograph index after encrypted that are not solved yet. In this paper, the judgment condition of view security, k- Anonymity was employed on the judgment for the security of cryptograph index. Two security of cryptograph index plan of constructing Hash functions and partitioning multi-sensitive attributes into buckets to establish cryptograph index were proposed aiming at the problem of information leaking by constructing cryptograph index. Finally, an optimal buckets partitioning strategy was proposed. The optimal query ratio can be gotten through buckets partitioning and cryptograph index creation under the condition of sub-interval number of buckets partitioning is limited which make the balance between security and efficiency of cryptograph index.