Multiple-keyword Ranked Symmetric Searchable Encryption Technology Research In Cloud Storage

In modern time,the data outsourced to the cloud has been widely used.However,with the application of the cloud storage,the security issue related to cloud storage is also gradually exposed.As well the privacy preserving of the users' data in the cloud has attracted the popularity's attention,which has been the bottleneck of restricting the further development of the cloud storage.Therefore,to guarantee the privacy of users' data,the data should be encrypted before outsourced to the cloud.But the traditional keyword searchable techniques based on the plaintext are not applicable to the cipher-text.Though searchable encryption can achieve cipher-text retrieval,the early searchable encryption schemes primarily support single keyword,and then to promote the retrieval accuracy,the schemes of supporting the multiple keywords ought to be proposed,and the server is able to return the most desired results as far as possible.As a consequence,how to achieve the multiple keywords ranked search based on the cipher-text securely and efficiently is becoming the hot issue in the academia and industry in the last fewer years.This paper analyzes the currently existing Cao's MRSE(Multi-keyword Ranked Search over Encrypted data),of which the first one makes use of k NN(k nearest neighbor)and the inner product matching to achieve the privacy of users' data,yet the security of the scheme is not imperial enough.The virtual keyword is introduced to extend vector in the second solution,which improves the security to some degree,but the returned results are not adequately accurate.This paper analyzes the advantages and disadvantages of Cao's schemes,and a modified searchable encryption scheme(MMRSE)supporting multi-keyword ranked is proposed.In terms of keywords,the scheme employs the vector space model to transform the keywords and documents to the vectors,then judges the similarity between them to achieve the multiple keywords search.Considering the retrieval efficiency,the scheme uses the TF-IDF weight model to rank the returned results.The scheme also adopts the CCA2 homomorphic encryption algorithm to ensure that the cipher of keywords is not easily destroyed by the attackers.