MBE: Client Side Secure Data Deduplication Based on Merkel Tree

Abstract

Several secure data deduplication protocols are recently proposed to maintain unique data copies at the cloud service providers (CSP’s) and utilize the storage space efficiently. Though many Message Locked Encryption (MLE) based protocols exist to deduplicate but delimits in handling various attack scenarios like poison attack, dictionary attack, side channel attack and frequency analysis attack. Moreover, maintaining the integrity of the data while being stored at the CSPs poses fresh challenges. In this thesis, we propose MBE, a client side data deduplication protocol that handles the aforementioned attack scenarios and also validates the data integrity at the client as well as the server end. Further, the MBE protocol authenticates the client through a proof of work (PoW) protocol based on the merkle tree and also detects the tamper at the block level. Communication and computation cost analysis are done for the proposed and the other state-of-the-art protocols to validate the efficacy of the proposed MBE protocol.