计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (1): 9-15.DOI: 10.3778/j.issn.1002-8331.1605-0170

• 热点与综述 • 上一篇    下一篇

Co-Work:基于HDFS的安全云存储数据一致性保持算法

林  穗,黄  健,姜文超   

  1. 广东工业大学 计算机学院,广州 510006
  • 出版日期:2017-01-01 发布日期:2017-01-10

Co-Work: Data consistency preserving algorithm based on HDFS in secure cloud storage system

LIN Sui, HUANG Jian, JIANG Wenchao   

  1. School of Computer Science and Technology, Guangdong University of Technology, Guangzhou 510006, China
  • Online:2017-01-01 Published:2017-01-10

摘要: 针对云存储平台中用户隐私和敏感数据的安全保护问题,在前期提出的基于自主可控机制的安全云存储模型ASOM(All Self-Organization Model)基础上,对MDSS端元数据的管理操作、MDSS和DMS之间的通信过程做进一步优化,针对ASOM模型实际场景引入锁思想实现两个节点一致性的Co-Work算法,完成MDSS和DMS节点之间的协同工作和数据的一致性保持。同时,考虑网络带宽对ASOM中读写效率的影响,引入随机表机制,改变DSS上报的时间结点,以提高ASOM整体读写效率。测试结果表明: 执行Co-Work算法后的ASOM模型实现了数据的物理与逻辑隔离,保证用户对元数据的自主控制和管理,而且随着数据尺寸增大读写效率明显提高,在数据达到1 GB时读写效率提高了12%。

关键词: 安全云存储, Hadoop分布式文件系统(HDFS), 一致性

Abstract: To guarantee the privacy protecting and sensitive data storage in cloud storage, this paper presents a consistency preserving algorithm, Co-Work, based on the actual scene and lock mechanism which can be implemented in ASOM(All Self-Organization Model). Co-Work algorithm can keep consistency between the metadata server and the storage device through a random map. Experimental results show that the ASOM model implemented Co-Work algorithm can isolate the physical data from logical data, avoid the autonomous users to control and manage the metadata, and increase the data transport efficiency. The read/write efficiency can be increased by 12% when the file size reachs 1 GB.

Key words: secure cloud storage, Hadoop Distribute File System(HDFS), consistency preserving