Research On Searchable Encryption Algorithm Based On Paillier Homomorphic Encryption

The data outsourcing model of cloud storage can provide high-quality data services,but it also brings risks to data security.Data encryption can protect the security and privacy of data in cloud storage,but it also brings about difficult problems of data retrieval.Searchable encryption is to retrieve encrypted data on cloud storage without decrypting all data,and to ensure efficiency and security as much as possible.However,the existing searchable encryption schemes are still inadequate.On the one hand,many schemes only support accurate retrieval,and cannot complete the retrieval when the user enters a spelling error,or the scheme can achieve simple keyword error correction,but the error correction result is not accurate enough to meet the requirements of searchable encryption environment.On the other hand,although most searchable encryption has been changed from single keyword to multi-keyword research,the research on multi-keyword sorted retrieval is still insufficient.And in order to realize the function of multi-keyword sorting retrieval,many schemes sacrifice the retrieval efficiency and security.In view of the above two problems,this paper conducted the following research:A keyword error correction algorithm based on word probability and edit distance is proposed.In this paper,on the one hand,the editing distance of different keywords is taken as one of the reference factors for error correction;on the other hand,a keyword probability library is introduced to take the word frequency of keywords appearing in the document as another reference factor for error correction.Through the comprehensive consideration of editing distance and word frequency,the accuracy of word error correction is further improved,so as to improve the precision and efficiency of ciphertext retrieval in cloud storage.This paper proposes a searchable encryption scheme based on Paillier homomorphic encryption.In this paper,Paillier homomorphic encryption algorithm is introduced into the scheme of searchable encryption,and the correlation score tf-idf value between keywords and documents is encrypted by Paillier algorithm.The homomorphism is utilized to realize the multi-keyword sorting retrieval function.At the same time,by introducing the queued forwarding mechanism of the agent,the connection between the keyword and the relevant document is cut off,and the connection between the retrieval keyword and the document is confused by adding "pseudo-retrieval request" in the retrieval process.Maximum security is guaranteed in the retrieval process.The experimental analysis shows that the word error correction algorithm in this paper can achieve more accurate error correction than the original edit distance,and can support multi-keyword correlation ranking retrieval.Compared with other recent research work,this scheme achieves more accurate and efficient retrieval.