Parallel FHE Bootstrapping Calculator for Scalable Ciphertext Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing FHE hardware accelerators face challenges with fixed parameter settings, limited scalability, and inefficiencies in bootstrapping operations, particularly when handling large data volumes and broader data ranges, leading to performance bottlenecks and security limitations.
Innovation Solution
A hardware design incorporating a memory, controller, task scheduler, and multiple processing units, capable of decomposing ciphertext into tasks and performing computations as Learning With Errors over Rings, with a bus for data transmission, allowing flexible configuration and parallel processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If FHE computations are performed on encrypted data using traditional encryption techniques, then data privacy protection is improved, but computational speed deteriorates
Solution Approach 1:
The patent segments the FHE computation process into multiple independent stages: key generation, encryption, homomorphic computation, and decryption. Each stage is handled by specialized hardware modules that can operate in parallel, thereby maintaining security while improving overall computational throughput
Solution Approach 2:
The patent introduces trusted execution environments (TEE) and secure enclaves as intermediary layers between the encrypted data and processing units. These intermediaries enable secure computation on encrypted data without requiring full decryption, thus maintaining privacy protection while reducing computational overhead
2Reliability
If FHE parameters are increased to support larger data ranges and higher security levels, then security and data range are improved, but device complexity and scalability deteriorate
Solution Approach 1:
The patent implements dynamic parameter configuration capabilities that allow the FHE system to adjust security parameters, polynomial ring dimensions, and modulus values based on specific application requirements. This dynamic adaptability enables the system to optimize between security levels and computational efficiency without requiring complete system redesign
Solution Approach 2:
The patent designs a universal FHE accelerator architecture that can handle multiple FHE schemes (e.g., BFV, CKKS, FHEW) and parameter configurations through a common hardware framework. This multi-functionality allows the same device to support both high-security applications with large parameter sizes and lower-security applications with smaller parameters
3Device complexity
If bootstrapping operations are performed using single-unit processors, then implementation simplicity is maintained, but computation time increases significantly
Solution Approach 1:
The patent segments the bootstrapping operation into multiple parallel sub-tasks that can be executed simultaneously by different processing units. The bootstrapping process is divided into keyswitch operations, polynomial multiplications, and noise management tasks, each handled by dedicated hardware modules
Solution Approach 2:
The patent merges multiple bootstrapping operations into a single unified processing pipeline, allowing concurrent execution of multiple ciphertext bootstrapping tasks. The hardware architecture combines keyswitch units, multiplication units, and addition units into an integrated system that processes multiple ciphertexts in parallel
4Reliability
If FHE is used to protect sensitive data during cloud computing, then data security is improved, but communication costs and memory requirements increase
Solution Approach 1:
The patent performs preliminary optimization of FHE parameters and pre-computes certain cryptographic components (such as evaluation keys and switching keys) before actual data processing. This preliminary preparation reduces the memory burden during actual encrypted computation by having essential data ready in advance
Data Source
AI summary
The accelerated bootstrapping fully homomorphic encryption calculator includes a memory, a controller, a task scheduler, a plurality of processing units, and a bus. The memory stores ciphertext structured as learning with errors over rings. The controller manages storing the ciphertext and generates instructions for bootstrapping. The task scheduler organizes tasks based on the controller's instructions, and the processing units decompose the ciphertext and perform computations to generate intermediate or final results. The bus connects all components and handles the transmission of ciphertext, instructions, and results.