Multi-Secret RLWE Key Switching for Faster Homomorphic Linear Transforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional homomorphic encryption methods face challenges in processing linear transformations and bootstrapping at high speed due to the time-consuming nature of operations, which hinders efficient processing of homomorphic cryptograms.
Innovation Solution
A device and method for parallel processing of linear key exchange of multiple homomorphic cryptograms using multi secret RLWE, involving packing and transforming cryptograms with a multi secret switching key to reduce bootstrapping time and improve performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional homomorphic encryption methods are used for linear transformation and bootstrapping, then security and encryption functionality are maintained, but processing time is excessive and speed is low
Solution Approach 1:
The patent segments the linear transformation process by separating the linear part from the remaining part of the cryptogram. This allows the linear part to be processed independently and in parallel with other operations, significantly reducing the sequential processing time bottleneck in conventional homomorphic encryption methods
Solution Approach 2:
The patent merges multiple processing steps into a unified approach where the linear transformation is integrated with the bootstrapping process. By combining these operations and processing them simultaneously where possible, the system reduces the total time required compared to performing them as separate sequential steps
Solution Approach 3:
The patent performs preliminary processing of the linear part of the cryptogram before the main bootstrapping operation. By pre-processing and preparing the linear components in advance, the system reduces the computational burden during the critical bootstrapping phase, thereby improving overall processing speed
2Productivity
If multiple homomorphic cryptograms are processed sequentially, then processing accuracy is maintained, but productivity is low
Solution Approach 1:
The patent divides multiple homomorphic cryptograms into separate processing streams that can be handled independently. Each cryptogram's linear part is extracted and processed in parallel with others, allowing the system to maintain processing accuracy while significantly increasing throughput through parallel execution
Solution Approach 2:
The patent creates a universal processing framework that handles multiple types of homomorphic cryptograms using the same parallel processing architecture. The system uses a unified approach for linear part extraction and transformation that works across different cryptogram formats, reducing the need for separate specialized processing paths
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computing device for linearly transforming multiple homomorphic cryptograms is disclosed. The computing device comprises a communication device for receiving an input of multiple homomorphic cryptograms, a memory for storing at least one instruction and the multiple homomorphic cryptograms, and a processor for executing the at least one instruction. The processor executes the at least one instruction to: pack the multiple homomorphic cryptograms by using multi-secret RLWE that share a linear part of the cryptograms; and perform key exchange of the packed multi-secret RLWE cryptograms by using a multi-secret switching key obtained in the format of multi secret RLWE.