An Efficient and Secure Deduplication Scheme Based on Rabin Fingerprinting in Cloud Storage

摘要

Data deduplication has been widely used in backups to save storage space and network bandwidth. In order to improve efficiency and security of the current deduplication schemes, this paper proposes a secure deduplication scheme. The proposed scheme supports cloud storage servers to eliminate deduplicate data before users' encryption operations, which can reduce computation overheads. Our scheme realizes variable-size block-level deduplication based on the technique of Rabin fingerprinting. Rabin fingerprinting selects blocks based on property of the block contents and hence supports data update and variable files. In the proposed scheme, a trusted third-party server is introduced to randomize the convergent keys and manage them. Security analysis indicates the proposed scheme is secure against offline brute-force dictionary attacks.