Dynamic Multi-Keyword Ranked Search Based on Bloom Filter Over Encrypted Cloud Data
Cloud computing has become a popular approach to manage personal data for the economic savings and management flexibility in recent year. However, the sensitive data must be encrypted before outsourcing to cloud servers for the consideration of privacy, which makes some traditional data utilization functions, such as the plaintext keyword search, impossible. To solve this problem, we present a multi-keyword ranked search scheme over encrypted cloud data supporting dynamic operations efficiently. Our scheme utilizes the vector space model combined with TF $\times $ IDF rule and cosine similarity measure to achieve a multi-keyword ranked search. However, traditional solutions have to suffer high computational costs. In order to achieve the sub-linear search time, our scheme introduces Bloom filter to build a search index tree. What is more, our scheme can support dynamic operation properly and effectively on the account of the property of the Bloom filter, which means that the updating cost of our scheme is lower than other schemes. We present our basic scheme first, which is secure under the known ciphertext model. Then, the enhanced scheme is presented later to guarantee security even under the known background model. The experiments on the real-world data set show that the performances of our proposed schemes are satisfactory.