Research On Privacy-preserving Searchable Encryption Scheme For Cloud Computing

The cloud server provides high-quality data storage services and reduces the data storage and maintenance overheads at the user side.In order to protect the data privacy,it is a common solution to encrypt data and then upload them into cloud server for outsourced storage.However,the unreadability of the ciphertext also hinders the data usability.Searchable encryption(SE)technology is an important approach to deal with this problem,which enables the users to search over encrypted data to realize effective data utilization.In this paper,the novel searchable encryption schemes are proposed according to the deficiency of the existing schemes.The main contents of these schemes are as follows.(1)In view of the problem that the existing searchable encryption scheme is not efficient and accurate,this paper proposes the basic fast multi-keyword semantic ranked search scheme(BFMSS)and enhanced fast multi-keyword semantic ranked search scheme(EFMSS).In the BFMSS scheme,Firstly,for the first time,the concept of weighted domain scoring is introduced to searchable encryption to reflect the weighted of the keywords in different domains.Secondly,combining the semantic similarity,the weighted domain score and the relevance scores,we design a three-factor ranking algorithm to obtain the more accurate document scores.To improve the efficiency of MRSE(multi-keyword ranked search over encrypted cloud data),we design vector block marking algorithm and mark,which effectively filters a large number of irrelevant documents.The computation overheads for calculating the relevance scores and ranking are greatly reduced.In the EFMSS scheme,we further design vector segmentation encryption algorithm to reduce the computation overhead of generating encrypted indices.The algorithm partitions the document index vector into several sub-vectors,which are encrypted by the matrices with smaller dimensions.Theoretical analysis and experimental results demonstrate that both the BFMSS scheme and EFMSS scheme proposed in this paper achieve the multi-keyword semantic ranked search with high efficiency.It improves the retrieval efficiency and returns more accurate ranking results in the BFMSS scheme.Meanwhile,it also reduces the encrypted index generation time in the EFMSS scheme.(2)In view of the problem that the existing Chinese keyword fuzzy searchable encryption schemes have to pre-define the fuzzy sets,which wastes a lot of storage and computation overheads.In this paper,we propose a novel basic Chinese multi-keyword fuzzy rank searchable encryption scheme(BCMS).First,we design Chinese keyword vector generation algorithm based on pinyin string to convert a pinyin string into a keyword vector.The algorithm converts the Chinese keyword to the pinyin string,which is partitioned based on the mandarin initials,finals and tones of pinyin.The pinyin string is converted into a keyword vector.Then,the locality-sensitive hashing(LSH)and Bloom filter are utilized to construct the fuzzy keyword search algorithm.Moreover,the original encrypted indexes of the stored files are not influenced when a file is added(or deleted)in the system,which realizes the dynamic update of the files.The theoretical analysis and experimental results indicate that the proposed BCMS scheme realizes Chinese multi-keyword fuzzy search without constructing a large fuzzy set,achieves a dynamic document update,and saves a large amount of storage and computation costs.(3)In view of the problem that the basic Chinese multi-keyword fuzzy rank searchable encryption scheme(BCMS)is not accurate,this paper proposes the enhanced Chinese multi-keyword fuzzy rank search scheme ?(ECMS-?)and enhanced Chinese multi-keyword fuzzy rank search scheme ?(ECMS-?).The ECMS-? and ECMS-? scheme utilize Chinese keyword vector generation algorithm 1 and algorithm 2 based on unigram to convert a pinyin string into a keyword vector,respectively.It enlarges the difference of diverse change patterns in the pinyin string to make the rank result more accurate in Chinese keyword vector generation algorithm 1.The ECMS-? scheme takes into consideration the word location information in the Chinese keyword vector generation algorithm 2 such that the permutation of the words in a keyword may result different syllable segmentation set,which is not realized in ECMS-? scheme.The theoretical analysis and experimental results indicate that the proposed ECMS-? and ECMS-? scheme realize Chinese multi-keyword fuzzy search,and return more accurate search result than the BCMS scheme.