The Research On Cloud Data Integrity Verification And Repair Scheme

Cloud computing services generally use a redundant backup method to ensure data security,which will copy multiple replicas while storing source data.When the source data is corrupted or lost,the cloud can repair the data with one of the replicas.However,for reducing overhead,some cloud service providers may do something that are not conducive to user data.For example,for increasing user capacity,they may change or delete replicas of the data which is not commonly used by users to free up storage space.Once the data is corrupted,it will be irreparable.Cloud data protection mainly includes detection and repair.In the detection part,the existing cloud data integrity verification scheme based on MHT(Merkle Hash Tree)structure takes one file as the verification unit.Multi-file verification in DFS(distributed file system)will lead to high communication complexity and low verification efficiency,and most of these schemes don't support multi-replica verification or dynamic operation of data.In the repair part,although the existing cloud data repair scheme based on LRC(Local Repair Code)requires fewer participating nodes when repairing one file block,the multi-block simultaneous repair still needs the participation of all nodes in the group,which can result in a large reading and communication overhead.In order to address these problems,this paper works as follows:(1)This paper proposes a coordinate-based MHT structure,which we called CMHT.The corresponding generation algorithm and dynamic updating algorithm are designed.The dynamic updating algorithm can achieve multi-node updating with minimum computation.In order to reduce the communication complexity and improve the efficiency of verification,a cloud data integrity verification scheme based on CMHT is proposed.The scheme constructs CMHT in datanodes,so it can support multiple-replica verification.This paper also designs corresponding protocols to support remote data verification,abnormal data location and data dynamic operation.(2)This paper improves LRC and proposes FLRC(Fully LRC),which has strong error-correcting ability and requires less overhead of reading and communication for multi-block repair,and we also designs the corresponding coding and decoding algorithm.Then,a data repair scheme based on FLRC is proposed,and the corresponding protocols are designed to support abnormal data repair and dynamic operation of data.(3)The above two schemes are simulated and analyzed in this paper.The results show that the verification schemes can greatly improve the efficiency of verification,and the repair schemes can effectively reduce the reading and communication overhead for multi-block repair.