| With the rapid development of cloud computing technology,more and more individuals and businesses choose to store their data on the cloud server,so as to save expensive data maintenance and management overhead.However,the attendant problem is that the data on cloud server grows rapidly,which leads to an increase in redundant data and a huge waste of cloud storage resources.Data deduplication technology,as a new technology to reduce redundant data and improve storage efficiency,has become a research hotspot at home and abroad.It allows the server to store a single data instance and delete duplicate copies,thereby reducing storage overhead and saving upload bandwidth.However,many data deduplication schemes mainly consider the efficiency of deduplication and the security of data,while ignoring the management of convergence key.The security,storage and management of the convergent key is also a problem that cannot be ignored.Secondly,most of the data integrity audit schemes use the authentication structure of Merkle hash tree to support the dynamic operation.But the major problem of Merkle hash tree is that it lacks a fast and efficient method to query leaf nodes.Moreover,due to the characteristics of the binary tree,the Merkle hash tree may become unbalanced after multiple update operations,which will lead to increased overhead for integrity verification and dynamic operation.In order to improve the efficiency of data integrity verification,a new authentication structure and data integrity verification scheme are urgently needed.Aiming to tackle these problems,the main contributions of this thesis are given as follows:1、This thesis proposed an efficient data deduplication and key management scheme.The scheme uses the key mask based on user identity to encrypt the convergence keys and upload them to the cloud server,which solves the storage and management problems of convergence keys.Then using the blind signature and third-party auditors to sign the key mask.Secondly,the scheme uses a combination of file level and file block level methods to deduplicate data.Compared with the file-level deduplication,it has lower granularity and better deduplication efficiency.In addition,the scheme also introduces a file management directory to manage the content of files and blocks,convergence key and the permissions of legitimate users.In the data integrity verification stage,aggregate signatures and bloom filters are used to improve the efficiency of data verification and tag query.Through theoretical analysis and experimental,it shows that our scheme has better performance.2、This thesis proposed an authentication structure of rank-based multi-branch tree.By combining multi-branch tree and linked list,we propose a new authentication structure.The depth of the authenticated structure is lower than that of Merkle hash tree,which effectively reduces the communication and computational overhead during integrity verification.In order to optimize the query efficiency,the rank is added to the nodes of the authentication structure tree.And according to this structure,a data integrity detection scheme which can supporting dynamic operation is proposed.The scheme introduces the third-party audit to perform integrity audit work,and adds random mask to protect data privacy.Through the theoretical analysis and experimental simulation,it is proved that the construction cost of the authentication structure is lower than that of other schemes,and the data integrity detection scheme based on this structure performs well in various performances. |
