Design And Implementation Of A Protocol For Proof Of File Ownership Based On Convergent Encryption

Today, with the booming development of cloud computing, cloud storage is widely applied in daily data storage services. In order to save network upload bandwidth and storage space,client-side deduplication is performed in the cloud storage system, which takes the hash value of the intended file as a compact representation of it. By comparing the hash values between the intended file and the original, the cloud storage server can determine whether the uploaded file is duplicated, and if so, the client does not need to upload the file.However, the mechanism may introduce a new security issue. An attacker who only needs to show some hash values of a file to the cloud storage server can successfully deceive the latter that he owns the file, hence obtain the entire file illegally.To implement the client-side deduplication, current deduplication schemes and proof of file ownership protocols are studied systematically. Their working mechanisms are analyzed,and problems about performance and security are also summarized in this thesis. Based on these preparations, a new protocol for proof of file ownership based on convergent encryption is proposed. With the cooperation of the client and the cloud storage server to generate the encryption key, it solves the problem that convergent encryption is vulnerable to offline dictionary attacks. It also improves the security by using the signatures of original file data blocks as the evidence of file ownership, and significantly reduces the communication overheads in the protocol interaction and the computational costs of cloud storage server by exploiting aggregate verification. Finally, based on the Hadoop distributed file system, a cloud storage server is built for distributed storage of user files and the location and extraction of data blocks. The interactions of our proof of file ownership is implemented on this cloud storage environment. Performance test and security analysis results show that the scheme is efficient, especially in reducing the computational costs of the cloud storage server, and the security of the scheme is better than other existing protocols.