Research On Searchable Encryption For Dynamic Data In Cloud Storage

With the maturity of Internet technology and the emergence of numerous applications,the scale of data has reached an unprecedented level.As a production factor,data contains an immeasurable value.At the same time,the rapid development of technology has promoted the maturity of cloud technology,and cloud servers are general in daily production and life.Research on using cloud technology to achieve secure data storage and utilization has become a hot issue.Cloud storage technology provides users with an economical and convenient data storage method,which attracts users to upload a large amount of data to the cloud.However,due to various attacks,frequent data leakage incidents disrupt the property and compromise the data security of enterprises and consumers.Encrypting and uploading data to the cloud has become a common way to protect data privacy.The explosive growth of data keeps the ciphertext data in the cloud in a dynamic state of constant addition and deletion.Therefore,it is necessary to study encryption techniques for dynamic data in cloud servers to achieve secure data usage.Since data encryption disrupts the functionality of data search and seriously hinders the use of data,searchable encryption,which can support ciphertext search,has become a notable scientific issue of common concern in academia and the industry.Generally speaking,users extract keywords from the data,upload an encrypted index that denotes the relationship between the two,and search for ciphertext by querying keywords.Currently,some searchable encryption schemes support data updates,such as additions and deletions,for the dynamic data scenarios of cloud storage.However,the research on searchable encryption in dynamic data scenarios still faces the following issues in practical applications.Firstly,searchable encryption that supports data updates focus on single keyword search,which cannot meet the complex semantic search needs of users in large-scale dynamic data;Secondly,in dynamic data scenarios,achieving data management through data integration on cloud servers and allowing to search the integration results as independent objects is a practical requirement and a foundation for more complex data management;Finally,it is vital to realize dynamic data sharing in cloud storage and retain users’ ability to search data in terms of promoting data flow and mining data value to solve the problem of data islands.However,the existing research on security still needs to be improved.So,based on the existing research on searchable encryption,this thesis takes dynamic data in cloud storage as the object to research complex semantic retrieval,data integration,and data sharing,and the main work is as follows:(1)Tree-based index wildcard searchable encryptionImplementing complex semantic encryption search in dynamic data is the crux of users to efficiently and securely search for data that meets specific queries in dynamic,massive data on the cloud.It is popular to achieve complex semantic retrieval based on wildcards due to their flexibility.However,existing research focuses on the efficiency of static data search,and is stuck on the problem of improving search speed and accuracy simultaneously.At the same time,there is no universal security model for wildcard searchable encryption.To solve the above problems,we design efficient and secure wildcard searchable encryption that supports data updates based on tree indexes,and define a general security model.Firstly,we enrich the extracted characters,and divide them into three types:A-type characters,AB-type characters and BF-type characters according to their contained information.Secondly,according to the order of the three types of characters,the "Three-step search method" is innovatively proposed and a tree-based index is designed from top to bottom.Thirdly,realize efficient and accurate wildcard search.Similarly,the proposed scheme supports users to add and delete data,and realizes complex semantic retrieval in dynamic data.On the basis of experimental tests and theoretical analysis,the proposed scheme is significantly better than the existing scheme in search speed and accuracy.In addition,we define a general security model for wildcard searchable encryption:the Non-Adaptive Indistinguishable Security Model for Chosen Character Set Attacks(IND-CCSA).The access characters set(AC)places no restrictions on characters and is instantiated by a matrix to describe the relationship among characters,keywords and files,which solves the problem that the existing security model is limited to a specific character set.The strict security certificate proves that the proposed scheme is IND-CCSA secure.(2)Mergeable Searchable encryption for keywordsTo realize the data integration of data on the cloud server,we propose to merge data by merging keywords due to the observation of using keyword search data in searchable encryption.Realizing dynamic data merge is a prerequisite for users to efficiently search for dynamic massive data in the cloud and an important way for users to safely manage data.Although users can use dynamic searchable encryption to achieve the goal of merging keywords through "Search-Merge Locally-Update",it will bring serious computing burden to users and cloud servers,and the risk of leaking what data is composed of the merge result.To solve these problems,we propose mergeable searchable encryption,that is,users can integrate data identified with different keywords(w1,...,wn)into a whole identified with the new keyword w and store it in the index,so that users can obtain the res ponding data by querying w and search w1,...,wn normally.Additionally,when there are updates for w1,...,wn,the data identified with the keyword w will update correspondingly and realize merging consistency.Specifically,a compressed homomorphic encrypted index is designed based on the super-increasing sequence,so that the merge operation is transformed into the addition operation of the corresponding index.The oblivious data retrieval designed based on the distributed point function guarantees that it will not disclose which keywords(w1,...,wn)are involved in w,named Parallel Privacy.When achieving merging consistency,we introduce a cloud server to outsource some computation of users,which reduces the computing burden of users.On the basis of experimental testing and security demonstration,the proposed scheme improves efficiency and realizes parallel privacy,and finally realizes the secure data integration of dynamic data in cloud storage.(3)Trapdoor updatable searchable encryption for multiuserFor the secure data sharing of dynamic data,we consider the classic scenario of multiuser,which involves the data privacy leakage caused by dynamic operations in the cloud and the changes in access rights to data for users,the basic to study more complex data sharing.In order to solve the above problems simultaneously,we construct a dynamic multiuser symmetric searchable encryption scheme based on homomorphic message authentication code(HomMAC),bitmap(Bitmap),and pseudorandom function(PRF).The index constructed by HomMAC and Bitmap ensures that the correspondence between files and trapdoors will not be leaked in the process of adding and deleting data,which solves the problem of efficiently and securely updating trapdoors,and realizes forward privacy and backward privacy.At the same time,by controlling the user’s ability to update trapdoors,the management of user access data rights is realized,ensuring that only authorized users can get the latest trapdoor and search data.On the basis of experimental testing and security demonstration,the proposed scheme improves efficiency,realizes forward privacy and backward privacy,and finally realizes secure dynamic data sharing in the cloud.In summary,this thesis aims to realize the secure use of dynamic data in cloud storage.It studies the technology of searchable encryption,especially the complex semantic retrieval,data integration,and data sharing problems faced in practical applications.We design corresponding solutions,improve some of the research in searchable encryption,and hope to contribute to ensuring data security and promoting the construction of digital China.