Homomorphic Encryption Search via Offline Preprocessing

The patent segments the computation process into two phases: an offline preprocessing phase where encrypted datasets are prepared and stored in a database, and an online query phase where search operations are performed. This segmentation allows computationally intensive operations to be performed in advance when resources are available, while queries execute quickly when needed, thus resolving the contradiction between security and computation time.

The patent performs preliminary actions by pre-processing encrypted datasets during an offline phase, including creating encrypted indexes and preparing data structures before they are stored in the database. This preliminary preparation reduces the computational burden during actual search operations, enabling fast query execution while maintaining security through homomorphic encryption.

The patent divides the system into a client component that performs lightweight encryption and query operations, and a server component that stores pre-processed encrypted data. This segmentation reduces the computational resource requirements on individual devices, as the heavy lifting is distributed to the server during offline preprocessing, while the client maintains low resource usage.

The patent creates encrypted copies of datasets and stores them in a database structure optimized for efficient retrieval. By pre-processing and storing encrypted versions of data in optimized formats, the system reduces the computational resources needed during actual search operations, as the server can perform operations on pre-processed encrypted data rather than raw encrypted data.