Homomorphic Encryption Bootstrapping Interval Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Homomorphic encryption systems face inefficiencies due to the need for repeated bootstrapping operations after multiple multiplications, which significantly increase execution time for complex computations.
Innovation Solution
A method and apparatus for data analysis that acquire a ciphertext for precomputation results from a client device and generate an encrypted computation result using homomorphic operations, allowing for efficient performance of remaining operations by dividing target data into parts and performing specific sub-algorithms, including multiplication and comparison operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiplication operations are repeatedly performed in homomorphic encryption, then the computation capability is improved, but the ciphertext becomes unsuitable for further homomorphic operations requiring bootstrapping which significantly increases execution time
Solution Approach 1:
The patent applies bootstrapping operations as preliminary actions at predetermined intervals during the computation process. Instead of waiting until the ciphertext becomes completely unsuitable for homomorphic operations, the system proactively performs bootstrapping at strategically chosen points to refresh the ciphertext and extend its usability for further multiplications, thereby reducing total execution time
2Reliability
If bootstrapping operations are performed frequently to maintain ciphertext suitability, then the reliability of continuous computation is improved, but the execution time significantly increases
Solution Approach 1:
The patent performs bootstrapping operations at predetermined intervals rather than continuously or excessively. This partial action approach maintains sufficient ciphertext suitability for homomorphic operations without over-processing, achieving a balance between reliability and execution time by applying bootstrapping only when necessary at strategic points in the computation
3Reliability
If all operations are performed on encrypted data using homomorphic encryption, then data confidentiality is improved, but the computation efficiency deteriorates due to frequent bootstrapping requirements
Solution Approach 1:
The patent segments the computation process into distinct phases: encrypted computation phases where homomorphic operations maintain data confidentiality, and intermediate bootstrapping phases that refresh the ciphertext. This segmentation allows the system to maintain security during computation while periodically addressing the efficiency degradation through structured bootstrapping intervals
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
An apparatus and method for data analysis are provided. The method for data analysis according to one embodiment includes acquiring, from a client device, a ciphertext for a precomputation result generated by applying some of a plurality of operations for performing an analysis algorithm based on target data to the target data; and generating an encrypted computation result for remaining operations of the plurality of operations by using the ciphertext.