Approximate Multi-keyword Rank Search on Encrypted Cloud Data

With the growing popularity of cloud computing, more and more data owners outsource their data to cloud storages because of the convenience for management. Since cloud storages are usually run by third party service providers, data security is obviously an important problem. A simple way to protect the security is to encrypt the data before outsourcing them to the cloud. Some existing techniques have studied keyword search over encrypted data. However, most of the techniques use a long index for the documents and do exact keyword search over the dataset. Those techniques may suffer from the computation and communication time as well as the storage space. In this paper, a keyword search over encrypted data scheme is proposed to improve the search performance in terms of computation time and required space. Essentially, the relations between the documents and selected keywords are first analyzed. Then, the principal information of the keyword-document relations is extracted to reduce the document indexes. The relevance of users' query and the documents is quantitatively evaluated to choose the final documents of interest. Experiments on real-world dataset show that the proposed scheme effectively improves the computation performance.