Research On Security Keyword Search Scheme In Cloud Environment

In the open cloud environment,massive data are outsourced to the cloud for use the computing and service resources of cloud,which can also save local storage and management costs.The data security utilization in the open cloud environment is an important prerequisite and basic guarantee for exerting data utility,realizing data sharing and promoting the development of digital economy.To protect sensitive information of outsourced data,data is stored in the cloud in encrypted form.How to realize keyword-based search on encrypted data is the first and key step of data security utilization.Secure keyword search technology,as one of the important technology for the encrypted data security utilization,has received extensive attention.In practical application,according to application scenarios,search requirements and security objectives,a specific security keyword search scheme is designed.The current research of security keyword search scheme in cloud environment is essentially a trade-off between security,search efficiency and search function.The existing secure keyword search schemes in cloud environment still have some problems,such as the similarity threshold d is not scalable for fuzzy keyword search,the intermediate result pattern will be leaked for forward privacy conjunctive keyword search,failed to achieve forward privacy and multi-keyword search at the same time,and the search results cannot be ranked or ranked inaccurate.In order to solve the above problems,this dissertation designs three security keyword search schemes.The main work and innovation of this dissertation can be summarized as the following three aspects:1.Designing a scalable secure fuzzy keyword ranked search scheme for cloud environment.Fuzzy keyword search can tolerate spelling errors and return all keywords whose morphology similarity is within the similarity threshold d.The existing secure fuzzy keyword search schemes can only support fuzzy search within the predefined threshold d,and with the increase of d,the storage cost will increase exponentially or the accuracy of search results will decrease.In this dissertation,we use edit distance as similarity measure and design an edit distance algorithm on ciphertext data,so that the proposed scheme can support fuzzy keyword search within any threshold d,the storage cost and search accuracy are not affected by d.The current secure fuzzy keyword rankable search schemes only take the keyword weight as the ranking criterion,ignoring the influence of keyword morphology similarity on ranking.In this dissertation,a two-factor ranking function is designed to rank the search results combining the keyword weight with the keyword morphology similarity,so that the ranking results are more suitable for the situation of fuzzy keyword search.2.Designing a secure conjunctive keyword search scheme with forward privacy.The security keyword search scheme in dynamic setting will leak forward information in the update process,which is threatened by file injection attack.The existing secure conjunctive keyword search scheme with forward privacy first executes single keyword search to find the file containing this keyword,and then determines whether the retrieved file contains all other sear-ch keywords.This two-stage search mode leads to the leakage of intermediate result pattern,and the computation and communication costs of the client are positively correlated with the number of search keywords.In this dissertation,a subset decision algorithm with long tail is designed.When the previous query vector is matched with the updated index vector,only ’False’ is output,thus achieve forward privacy.Based on this algorithm,this dissertation proposes a forward privacy secure conjunctive keyword search scheme without leaking intermediate result pattern,and has constant bandwidth consumption.At the same time,in view of the problem that the existing forward privacy secure keyword search scheme does not support the results ranking,this dissertation designs a ranking method based on TF × IDF weight,which realizes the results ranking in a dynamic setting,so that the clients can quickly locate the required files.3.Design a secure multi-keyword ranked search scheme with forward privacy.Multi-keyword search is a basic search requirement,which is suitable for the scene where multiple search conditions are query simultaneously.The existing security keyword search scheme with forward privacy in dynamic setting only realizes single keyword search and conjunctive keyword search,and does not realize multi-keyword search.This dissertation designs an easy-toimplement extended MRSE algorithm to realize secure multi-keyword search with forward privacy in dynamic setting.In the extended MRSE algorithm,the query and index is transformed as vectors,and carefully designed elements are added to the vectors to expand it.In order to enable users to quickly and accurately find the desired files,this dissertation designs a result ranking method based on the number of matching keywords and TF × IDF weight under dynamic setting.In addition,the proposed scheme only uses symmetric encryption primitives,so it has high search efficiency.Through security analysis and experimental evaluation on COVID-19 dataset,it is shown that the proposed scheme is secure and efficient.