Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

TFHE-SBC: Software Designs for Fully Homomorphic Encryption over the Torus on Single Board Computers

Marin Matsumoto, Ai Nozaki, Hideki Takase, Masato Oguchi

TL;DR

This paper tackles the practicality of fully homomorphic encryption on resource-constrained SBCs by addressing the inefficiency of TLWE-based TFHE on such devices. The authors propose TFHE-SBC, an SBC-focused framework that runs TRLWE encryption on the client and converts to TLWE on the server via a Sample Extraction protocol, complemented by a device library (fast RNG with seed expansion and Ziggurat Gaussian sampling) and an adapter library (TRLWE-to-TLWE conversion) that together reduce computation, communication, and energy costs. Key contributions include the first SBC-optimized TFHE client framework, ciphertext switching without additional noise, and memory reuse to minimize FFT overhead, achieving up to 2486× encryption speedups, 512× communication reductions, and 12–2004× energy improvements in experiments on a Raspberry Pi Zero 2W. These results demonstrate the practical viability of TFHE-based secure analytics on IoT data, with the current work focused on private-key encryption and future work needed to extend to public-key scenarios.

Abstract

Fully homomorphic encryption (FHE) is a technique that enables statistical processing and machine learning while protecting data, including sensitive information collected by single board computers (SBCs), on a cloud server. Among FHE schemes, the TFHE scheme is capable of homomorphic NAND operations and, unlike other FHE schemes, can perform various operations such as minimum, maximum, and comparison. However, TFHE requires Torus Learning With Error (TLWE) encryption, which encrypts one bit at a time, leading to less efficient encryption and larger ciphertext size compared to other schemes. Additionally, SBCs have a limited number of hardware accelerators compared to servers, making it challenging to achieve the same level of optimization as on servers. In this study, we propose a novel SBC-specific design, \textsf{TFHE-SBC}, to accelerate client-side TFHE operations and enhance communication and energy efficiency. Experimental results demonstrate that \textsf{TFHE-SBC} encryption is up to 2486 times faster, improves communication efficiency by 512 times, and achieves 12 to 2004 times greater energy efficiency than the state-of-the-art.

TFHE-SBC: Software Designs for Fully Homomorphic Encryption over the Torus on Single Board Computers

TL;DR

This paper tackles the practicality of fully homomorphic encryption on resource-constrained SBCs by addressing the inefficiency of TLWE-based TFHE on such devices. The authors propose TFHE-SBC, an SBC-focused framework that runs TRLWE encryption on the client and converts to TLWE on the server via a Sample Extraction protocol, complemented by a device library (fast RNG with seed expansion and Ziggurat Gaussian sampling) and an adapter library (TRLWE-to-TLWE conversion) that together reduce computation, communication, and energy costs. Key contributions include the first SBC-optimized TFHE client framework, ciphertext switching without additional noise, and memory reuse to minimize FFT overhead, achieving up to 2486× encryption speedups, 512× communication reductions, and 12–2004× energy improvements in experiments on a Raspberry Pi Zero 2W. These results demonstrate the practical viability of TFHE-based secure analytics on IoT data, with the current work focused on private-key encryption and future work needed to extend to public-key scenarios.

Abstract

Fully homomorphic encryption (FHE) is a technique that enables statistical processing and machine learning while protecting data, including sensitive information collected by single board computers (SBCs), on a cloud server. Among FHE schemes, the TFHE scheme is capable of homomorphic NAND operations and, unlike other FHE schemes, can perform various operations such as minimum, maximum, and comparison. However, TFHE requires Torus Learning With Error (TLWE) encryption, which encrypts one bit at a time, leading to less efficient encryption and larger ciphertext size compared to other schemes. Additionally, SBCs have a limited number of hardware accelerators compared to servers, making it challenging to achieve the same level of optimization as on servers. In this study, we propose a novel SBC-specific design, \textsf{TFHE-SBC}, to accelerate client-side TFHE operations and enhance communication and energy efficiency. Experimental results demonstrate that \textsf{TFHE-SBC} encryption is up to 2486 times faster, improves communication efficiency by 512 times, and achieves 12 to 2004 times greater energy efficiency than the state-of-the-art.

Paper Structure

This paper contains 20 sections, 2 equations, 4 figures, 7 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. General LWE
  4. TFHE Scheme
  5. Multiplying Polynomials via the Discrete Weighted Transform
  6. Related Work
  7. System Model and Security Assumptions
  8. Baseline Implementation
  9. Threat Model and Security Assumptions
  10. TFHE Performance Analysis and Challenges
  11. Comparative Analysis of TLWE and TRLWE
  12. Bottlenecks in TRLWE Implementation
  13. Applicability of GPUs to Polynomial Multiplication
  14. TFHE-SBC
  15. TRLWE to TLWE Conversion Protocol
  16. ...and 5 more sections

Figures (4)

  • Figure 1: End-to-end TFHE deployment solution by TFHE-SBC.
  • Figure 2: Architecture of TFHE-SBC: System workflow showing SBC-side TRLWE encryption and server-side conversion to TLWE ciphertext. The implementation features accelerated random number generation and single-execution FFT for secret keys.
  • Figure 3: Performance of TRLWEtoTLWEs. Results confirm TFHE-SBC's end-to-end encryption outperforms direct TLWE encryption, even with conversion overhead.
  • Figure : Sample Extraction: TRLWEtoTLWEschillotti2020tfhe