云存储数据验证算法的相关研究

发布时间：2018-06-03 18:40

本文选题：云存储 + 数据验证　；参考：《浙江大学》2013年硕士论文

【摘要】：随着云计算的广泛应用,越来越多的用户将自己的数据和文件存储在云服务器上。云存储在为用户带来便捷的同时,也带来一些问题。用户的数据和文件在服务器上可能会出现意外的篡改或丢失,而出现这种情况的时候用户和云服务器可能都发现不了。因此,需要定期对用户的数据进行验证来保证用户的数据和文件的完整性和正确性。本文首先对云存储数据验证的方法做了综述,介绍了数据验证的概念和验证方法的设计原则,并简要介绍了现有的方法。然后针对现有方法中存在的不足,提出了自己的改进思路和方法。现有验证方法,每次验证的计算过程只能对一个用户的数据进行验证,而本文提出一种方法,可以在一次验证的计算过程中,对多个用户的数据进行合并验证,不仅节省了验证的次数,更减少了验证过程中验证方与云存储方之间的数据传输量。然后针对用户的数据会动态变化的问题,已有的验证方法用一种叫做梅克尔哈希树(Merkle Hash Tree)的哈希树来维护文件信息,每次数据变化的时候对树的节点重新计算,对所有的节点都同样对待,而本文提出,在实际应用中,文件的使用频率是有差别的,应该根据文件使用频率来维护哈希树的结构,提高频繁使用的文件的优先级,并提出了相应的解决方法,通过模型和实验证明了本文提出的方法能够有效减少在文件动态变化时的计算量,提高了效率。
[Abstract]:With the wide application of cloud computing, more and more users store their data and files on cloud servers. Cloud storage brings convenience to users, but also brings some problems. Users' data and files may be accidentally tampered with or lost on the server, and neither the user nor the cloud server will find them. Therefore, it is necessary to verify the user's data periodically to ensure the integrity and correctness of the user's data and files. This paper first summarizes the methods of cloud storage data verification, introduces the concept of data validation and the design principles of verification methods, and briefly introduces the existing methods. Then, aiming at the shortcomings of the existing methods, the author puts forward his own improvement ideas and methods. The existing verification methods can only validate the data of one user in each verification process, and this paper proposes a method to merge and verify the data of multiple users in the calculation process of one verification. Not only the times of verification are saved, but also the amount of data transmission between verifier and cloud storage is reduced. Then, aiming at the problem that the user's data will change dynamically, the existing verification methods use a hash tree called Merkle Hash tree to maintain file information, and recalculate the nodes of the tree every time the data changes. All nodes are treated the same, but this paper proposes that in practical applications, the frequency of file use is different, the structure of hash tree should be maintained according to the frequency of file use, and the priority of frequently used files should be raised. The corresponding solutions are put forward. The model and experiments show that the proposed method can effectively reduce the computational complexity and improve the efficiency when the file changes dynamically.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP333

【参考文献】