FHE Accelerator Pipeline for Noise-Aware Encrypted Processing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing fully homomorphic encryption (FHE) systems face challenges in efficiently processing encrypted data due to noise growth, ciphertext expansion, and high computational overhead, particularly in operations like multiplications and permutations, which require expensive bootstrapping and key switching, limiting the number of homomorphic operations that can be performed before decryption failure.

Innovation Solution

A hardware accelerator for FHE is designed with pipelined stages, multiplexers, and specialized processing elements (PEs) to manage noise and reduce ciphertext expansion, utilizing conflict-free memory access and on-the-fly key generation to optimize operations like NTT and permutations, enabling efficient parallel processing of encrypted data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If fully homomorphic encryption is used to process encrypted data, then data security is improved, but computational overhead increases significantly

Engineering Contradiction:
Improvedata securityVSAvoidcomputational overhead
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The computational process is divided into multiple pipeline stages, with each stage handling specific operations (multiplication, addition, permutation) on encrypted data. This segmentation allows parallel processing and reduces the computational burden on any single processing element while maintaining security through continuous encryption throughout the pipeline.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces specialized hardware components (processing elements, memory units, pipeline registers) as intermediaries that operate on encrypted data without requiring decryption. These intermediaries enable efficient computation while preserving data confidentiality, bridging the gap between security requirements and computational efficiency.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Productivity

If multiple homomorphic operations are performed, then processing capability is improved, but noise growth increases leading to decryption failure

Engineering Contradiction:
Improveprocessing capabilityVSAvoiddecryption success
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The pipeline architecture enables continuous processing of encrypted data through multiple stages without interruption or decryption. Each stage processes data in an encrypted state, maintaining the cryptographic protection throughout the entire computation chain and allowing sustained processing capability while managing noise accumulation through systematic operation.

Inventive Principle:
Principle #20Continuity of useful action

Solution Approach 2:

The system performs preliminary noise management operations within each pipeline stage, including optimized multiplication and addition operations that control noise growth. By managing noise incrementally at each stage rather than allowing cumulative growth, the system extends the number of viable homomorphic operations before decryption.

Inventive Principle:
Principle #10Preliminary action

3Device complexity

If traditional processing architectures are used, then device simplicity is maintained, but processing efficiency of encrypted data deteriorates

Engineering Contradiction:
Improvearchitecture simplicityVSAvoidprocessing efficiency
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The patent implements a dynamic pipeline architecture where data flows through multiple specialized stages, each optimized for specific cryptographic operations. This dynamic structure adapts to different computational requirements while maintaining a relatively simple overall design, improving processing efficiency without excessive complexity through structured data flow and specialized processing elements.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS12547412B2Fully homomorphic encrypted processing acceleration
Publication Date: 2026.02.10 NIOBIUM MICROSYSTEMS INC
  • US12547412B2 patent drawing
  • US12547412B2 patent drawing
  • US12547412B2 patent drawing

AI summary

A device for processing fully homogeneous encrypted data comprises a command input with pipeline stages and a register file coupled to the command input via a non-pipelined stage and an ultimate pipelined stage. Further, the device includes various multiplexers to allow for usage of a multiplier for different command functions. Moreover, the device includes an adder element with a first input coupled to multiplexer outputs, a select line coupled to a penultimate pipeline stage of the command input, and an output coupled to the register file. A command on the command line passes through the pipeline stages as select inputs to the multiplexers to control the function of the multiplexers at that time.