Enhancing Privacy and Efficiency in Multi-Keyword Search on Encrypted Documents through Server-Side Index Computation
Vemula Sridhar, Ram Mohan Rao Kovvur, Marneni Dyna, Pelleti Swetha, Methuku Madhuri · 2024
Privacy-preserving Multi-Keyword Search has been extensively studied, with several contributions proposed in the literature. Traditional approaches involve creating an index on documents before encryption and uploading, with the encrypted index also uploaded to the server. A significant drawback of this method is that updating documents and the consequent index update require regenerating the index on all documents at the client-side and transferring them to the server, which hampers the system’s performance. This article introduces a novel approach to Privacy-preserving Multi-Keyword Search, where the index is calculated on encrypted documents at the server-side instead of the client-side, enhancing the system’s adaptability to document changes. To implement this, we retrieve and encrypt the desired keywords, then compress them before transmission to the server. Upon receiving the encrypted keywords, the server constructs an index that facilitates efficient document retrieval. This index enables the server application to generate a list of top N documents corresponding to the provided encrypted keywords. The original files are securely stored in CryptDB, ensuring access to only authorized clients. Our solution offers a secure and efficient alternative for Privacy-preserving Multi-Keyword Search, maintaining data integrity and confidentiality.