面向云存储的高性能数据隐私保护方法

High-performance Data Privacy Protection for Cloud

随着云计算和云存储技术的飞速发展,越来越多的企业和个人使用云存储来保存数据或备份数据.但用户将私有数据上传到云端的同时,也丧失了对数据的绝对控制权,用户数据的隐私保护问题成为云存储发展不得不解决的问题.为了解决这一问题,提出一种新的针对云存储的数据隐私保护方法BSBC（Bit Split Bit Combine）,它在上传前,将数据按照比特位进行拆分,重新组装后形成多个数据文件,再分别上传到云存储服务器;下载时,先将所有数据文件下载,然后通过位合并再恢复成原始文件.实验证明这种方法可以保护用户数据的隐私,同时可比传统加解密获得17～35倍的性能提升.然后将核心的位拆分、位合并代码模块用汇编语言进行优化,对汇编语言进行指令调度优化,以减少数据冲突和流水线停顿.最终,采用BSBC方法比传统加解密可以获得25～35倍的性能提升.

With the rapid development of technology of cloud computing and cloud storage,more and more businesses and individuals use cloud storage to store data or backup data.When uploading private data to the cloud,the user will lose the absolute control of the data,them data privacy protection becomes a problem that cloud storage has to solve.In order to solve this problem,BSBC （Bit Split Bit Combine）,a new data privacy protection method was presented.Before uploading the data,BSBC splits the data according to bit and re-assembled to form a number of data files,then uploads the data to cloud storage servers; when downloading the data,BSBC downloads all the data files,then through the hit combination,revert them to the original file.Experiments show that this method can protect the privacy of users＇ data,obtain 17～35 times performance improvement compared with traditional encryption.Then assembly language is used is used to optimize the core codes of bit split and bit combination,and instruction scheduling optimization of assembly language to reduce the data conflict and pipeline stalls.Eventually,compared with traditional encryption,BSBC can get 25 ～35 thmes performance improvement.