Research On Attribute-based Multifunctionalsearchable Encryption Scheme

In recent years,cloud storage has received extensive research and attention due to its low cost and high scalability.More and more companies and individuals choose to encrypt data and upload it to cloud storage.Although the confidentiality of data can be protected,it brings certain challenges to the retrieval and sharing of ciphertext data.How to achieve "one-to-many" fine-grained access control to the cloud ciphertext data and retrieve effective files from the huge ciphertext database has become an urgent problem to be solved.In addition,cloud services are semi-trusted.In an on-demand transaction and paid cloud service environment,cloud servers may act dishonestly in order to defraud service fees.Users may claim that they did not achieve the desired result and refuse to pay the service fee after receiving the target result.These will lead to the unfair phenomenon of "service-payment".Therefore,in response to the above problems,this thesis focuses on attribute-based searchable encryption schemes.The main contributions are as follows:(1)An attribute-based encryption scheme with revocation and ranking search functions for cloud storage is constructed.The scheme introduces an encryption server and a decryption server,which are respectively responsible for preliminary encryption and pre-decryption.In addition,in the user private key generation phase,the user private key is divided into two parts,and the private key component related to the attribute is sent to the decryption server.When the attribute revocation occurs,the decryption server is responsible for updating the attribute private key component.When the trapdoor submitted by the user is successfully matched with the keyword index,the scheme adopts TF×IDF rules to calculate the correlation score between the query keyword and the file,and according to the score,only the top-k file that best meets the user’s request is returned,which saves the user’s resource consumption.Performance analysis shows that the scheme has obvious advantages compared with existing schemes in computing and storage costs.In addition,under the q-DBDHE assumption and the general bilinear group model,it is proved that the scheme achieves IND-s CP-CPA security and IND-CKA security.(2)An attribute-based ranked multi-keyword searchable encryption with hiding access policy for smart health system is constructed.The scheme uses the vector space model to generate standard index vectors and query vectors,and calculates the correlation scores between the query keyword set and the file,and then realizes the ranking function.The blockchain technology is used to effectively solve the problem that the cloud server is semi-trusted and may return incorrect results.Two smart contracts are introduced,where the search smart contract checks whether the trapdoor matches the keyword index,thereby reducing the search cost of the cloud server.And the verifiable smart contract judges the legitimacy of the user and the correctness of the search results returned by the cloud.In addition,a random number is selected to blind the attribute information in the access policy,which protects the anonymity of the user.To reduce the computational overhead of the user in the decryption phase,the scheme introduces the decrypt-cloud service assistant to perform the initial decryption.And the performance analysis shows that the scheme has advantages compared with existing schemes in terms of function,computing and storage cost.The security analysis shows the security of the scheme.