Research And Implementation Of Data Integrity Verification Mechanism In Cloud Storage

Cloud storage technology makes the storage service more convenient. As aservice business mode with high practical value and market value, more and moreattention has been paid on it. The idea and concept of cloud storage service areattractive to both enterprise users and individual users. But as same as the other storageservice modes, the issue of data security in it is also users’ attention and concern. Inaddition, the technical characteristics of the cloud storage make the data securityproblem more complex than the traditional modes. Therefore, in the process of thedevelopment of cloud storage from concept to application, how to protect the securityof data in the cloud storage is an important subject.This thesis mainly researches on the data integrity verification mechanism incloud storage. The data integrity verification mechanism is divided into two main areas:remote integrity verification and general integrity verification. The prior one useschallenge/response model to verify the data integrity with probability and the latter isaccess-based that uses authenticated dictionary to verify the integrity of data which isaccessed.This thesis focuses on the research of general integrity verification mechanism.The three-party model of cloud storage is analyzed and two new data integrityverification mechanisms are proposed based on the existing related researches. Themain content of this thesis includes:(1) A data integrity verification scheme based on B+tree is proposed. Thisscheme achieves low time complexity in the verification, and supports flexible dataupdate operation. The application of the one-way accumulator makes it more efficientto change the digest of nodes during data update.(2) A data integrity verification scheme based on all-node stored tree is proposed.The size balanced tree (SBT) is used here to design the scheme. This structure ofall-node stored tree saves the storage space of the authenticated dictionary and it alsohas the equilibrium and efficiency as same as the balanced binary search tree. It is abrand new way to construct authenticated dictionary. (3) A prototype system of cloud storage is built based on the hadoop distributedfile system. This system supports data upload, data download and data updateoperation and provides platform support for the proposed integrity verification scheme.In addition, cryptography methods are used to authenticate B+tree and SBT structure.