CryptDB密文数据库系统并行方案研究

Investigation on Parallel Scheme of CryptDB Encrypted Database System

加密数据库对隐私数据的加密存储保护是解决当前互联网中用户隐私数据泄露的一种可行方案。鉴于互联网中每日产生的用户隐私数据规模巨大,传统的串行计算会导致隐私数据的加密存储时间消耗较长。为提高加密存储等数据处理的速度,将MapReduce并行框架与Crypt DB密文数据库系统进行有机结合,设计并实现了Crypt DB密文数据库系统并行加密和分布式存储的方案。并行方案采用任务调度算法、文件分割算法来提高其并行性和可控性,通过重写MapReduce框架中的Map方法来实现Crypt DB密文数据库系统的并行加密和分布式存储。基于由1个Master节点、3个Crypt DB节点和3个My SQL服务器构成的实验平台,进行了并行方案的实验验证及其性能分析。实验结果表明,所构建的并行方案在3个Crypt DB节点集群中的加速比可达到2.51,加密和存储时间节省了60.2%,可用于大规模关系数据的加密存储。

Encrypting data with encrypted DBMS is a feasible way to protect privacy of customer＇s sensitive data on the Internet.Due to the massive privacy data generated by Internet every day,the traditional serial calculation can lead to longer time consumption of encrypted storage of privacy data.Combined the characteristic of MapReduce and Crypt DB,a parallel insert and distributed storage scheme is designed and realized to improve the accelerate the speed of encrypting.Another two algorithms named Job Control and File Split are also proposed to improve the parallelism and controllability of this scheme.Map method in MapReduce is rewritten to achieve the parallel encryption and distributed storage of Crypt DB system.After doing the experiments and performance analysis on the platform consists of 1 Master node and 3 Crypt BD nodes and 3 My SQL server nodes,the experimental results showthat the speed-up radio of this proposed scheme can reach 2.51,and the total time cost of Cyrpt DB parallel scheme can be reduced to 39.8% on the cluster consisting of 3 Cyrpt DB nodes.It can be used into the encryption and storage of large-scale relational data.