Bin Encoding: A User-Centric Secure Full-Text Searching Scheme for the Cloud

Permitting users to search encrypted documents presents cloud storage providers with interesting challenges. Existing solutions target large corporations rather than individual users of the cloud. In order to serve all users, we propose a way of shifting most of the computational complexity from the client to the cloud by building and managing the index there, while ensuring that only the client can access the plaintext. This allows more sophisticated indexing and search ranking schemes to be implemented, including approximate search with multiple errors. Our method uses a many-to-one encoding scheme called "Bin Encoding", and this paper analyses its cryptographic strength against letter-frequency and dictionary attacks.