Abstract:

With the rise of cloud computing, data owners are increasingly outsourcing their complex data management systems to commercial public clouds to gain flexibility and cost efficiency. However, to protect data privacy, sensitive information must be encrypted before outsourcing, rendering traditional data retrieval methods based on plaintext keyword search ineffective. Therefore, enabling secure and efficient search services for encrypted cloud data has become a critical need. This paper addresses the challenge of privacy-preserving multi-keyword ranked search over encrypted cloud data (MRSE), a problem that has not been sufficiently explored in existing literature, which typically focuses on single-keyword or Boolean keyword search without differentiating search results. For the first time, we define the MRSE problem and propose a novel solution that meets stringent privacy requirements. We adopt the “coordinate matching” principle, which maximizes the number of matches between the query and documents, and use the “inner product similarity” metric to quantitatively measure this similarity. We introduce a basic MRSE scheme that leverages secure inner product computation and enhance it to meet privacy and efficiency needs under various threat models. Through comprehensive privacy and efficiency analyses and experiments on real-world datasets, we demonstrate that our proposed schemes offer a low overhead in both computation and communication, making them suitable for practical applications in cloud data search systems.

Description:

.