Research On Ciphertext-policy Attribute-based Searchable Encryption

As a new information technology,cloud computing has powerful distributed storage capacity and efficient data processing capacity.Data owners can outsource their data and compution task to cloud sever,and data users may search the files from cloud sever.While the cloud sever is semi-honest and curious,which threatens the security of data,there needs to encrypt data.How to search on ciphertext is a new problem.Attribute-based searchable encryption can not only protect data privacy,but also search with fine-grained access control on ciphertext.And only the user whose attribute satisfied access policy can decrypt successfully.This paper focuses on the problem in existing schemes: how to verify the integrity of ciphertext,enhance the efficiency of search and decryption,strong the security of scheme and deduplicate on ciphertext.We finished the following works:1.This paper proposed a synonym ciphertext-policy attribute-based searchable encryption scheme under bilinear groups of composite order.The proposed scheme achieves search quickly using TF ×IDF and ranking keywords and their synonyms.Through inserting the prediction value of file correlation in TF ×IDF method,cloud sever only returns the first K most related file.Through judging whether the user's attribute set satisfies access policy inserted in ciphertext,cloud transformation sever verifies the integrity of ciphertext and pre-decrypts ciphertext,which reduces decryption burden of user.In random oracle model,the proposed scheme uses double-system encryption to achieve full security based on three static assumption of composite order bilinear groups.2.This paper proposed a supporting block-deduplication searchable encryption scheme under large universe.This scheme uses hybrid cloud model to achieve data deduplication and searchable encryption simultaneously from ciphertext policy attribute-based encryption.Date owner splites data files into block and encrypts block files,then inserts ciphertext block into bloom filter tree.Through judging whether the linear secret sharing value embeded in the ciphertext of cloud and the ciphertext of bloom filter tree are included mutually,and matching tags quickly in bloom filter tree,the private cloud can perform fine-grained deduplication efficiently.Public cloud can search efficiently using homomorphic searchable method and keywords matching in bloom filter tree.In random oracle model,the proposed scheme can achieve privacy under chosen distribution attacks block-level secure deduplication and match-concealing searchable security based on Decisional Linear harden assumption and Decisional Bilinear Diffie-Hellman harden assumption.