Abstract: Deduplication is an efficient data compression and storage optimization technology in cloud storage systems.It can reduce storage space and transmission bandwidth consumption by detecting and eliminating redundant data.The convergence encryption adopted by existing cloud storage systems is vulnerable to brute-force attacks and the time cost of ciphertext generation is excessive.In this paper,an efficient deduplication scheme based on Merkle hash tree in hybrid cloud environment was proposed.The tag used to detect duplicated data is calculated by introducing privilege level function and label coefficients which can realize a secure deduplication system with different privilege levels.At the same time,an additional encryption algorithm is implemented,and cryptographic keys are generated by a Merkle hash tree.These keys are used to encrypt the plaintext at a file-level and block-level deduplication which ensures that the ciphertext becomes unpredictable.The security analysis shows that this scheme can effectively resist the brute-force attacks from internal and external attackers,and improve the confidentiality of data.The simulation results show that the proposed MTHDedup scheme can effectively reduce the computation overhead of ciphertext generation and the storage space of cryptographic keys.With the increase of the number of privilege sets,the performance advantage of MTHDedup scheme is more obvious.

Key words: Access control, Brute-force attacks, Data deduplication, Hybrid cloud storage, Merkle hash tree