• •    

信息存储技术学术会议+41+基于安卓平台的多云存储系统

孔琰1,赵帅兵1,刘若琳1,梁爽1,庄园1,冯世舫1,王刚2,刘晓光3,3,李忠伟1   

  1. 1. 南开大学计算机与控制工程学院
    2. 南开大学并行与分布式实验室
    3. 南开大学
  • 收稿日期:2016-11-18 修回日期:2016-11-22 发布日期:2016-11-22
  • 通讯作者: 孔琰

NCIS+41+ A Multi-Cloud Storage System based on Android

  • Received:2016-11-18 Revised:2016-11-22 Online:2016-11-22
  • Contact: Yan Kong

摘要: 针对云存储中存在的一系列的安全问题和移动互联网的完善和智能手机的兴起。本文提出了一个基于安卓平台的多云存储系统。首先,系统采用了加密机制来保证数据的机密性,然后切片数据,使用纠删码编码以实现容错能力,最后将多个数据块分散存储到不同的云存储服务提供商来解决厂商锁定的问题,确保服务的可用性。对于多服务商的安全认证问题,设计了基于加盐方法的令牌本地保护机制。针对安卓设备计算能力较弱的问题,实现了一种多级安全机制,利用不同加密、编码方式的组合,满足不同用户对于安全性和数据处理速率的权衡。并对关键操作如加密、编码进行了SIMD并行优化,进一步提高读写性能。实验结果表明,安卓多云存储系统实现了用户数据的云端安全存储,并提供了满意的读写吞吐率。

关键词: 多云存储, 加密, 纠删码, SIMD并行, 安卓

Abstract: Abstract: Because of a series of security problems in the cloud storage and the improvement of mobile Internet and the rise of the smart phone. This paper designed a multi-cloud storage system based on the Android platform. Firstly, this system adopted data encryption process to ensure the confidentiality of data, then sliceed the data, using erasure code to achieve fault tolerance aiming at ensuring data integrity, and finally a plurality of pieces of data was separated to a different cloud storage service providers on solving the problem of vendor lock-in, ensuring service availability. For the problem of multiple service provider security certification, this system designed a token protection mechanism based on local salt method. Because of weak computing capacity on Android equipment, it also designed a multi-level security mechanism, using the combination of different encryption and encoding methods to meet the users’ different needs of a balance between data processing speed and security. This system used SIMD technology for key operations such as encoding and encryption algorithm with parallel optimization for a better performance of reading and writing. The experimental results showed that the multi-cloud storage system based on Android achieves security cloud storage for users’ data, and provided a satisfactory rate of read and write throughput.

Key words: Multi-Cloud storage, Encryption, Erasure code, SIMD parallel optimization, Android

中图分类号: