Reliability Analysis of Fully Homomorphic Encryption Systems Under Memory Faults

2025-09-24

AI Summary

Full homomorphic encryption (FHE) systems face uncharacterized reliability risks under memory faults, yet no prior work has systematically evaluated their fault sensitivity or resilience. Method: We conduct the first comprehensive reliability assessment of FHE across major schemes—BFV, CKKS, and TFHE—using controlled fault injection experiments and analytical reliability modeling to quantify bit-flip sensitivity at both primitive-operation and end-to-end application levels. We further compare conventional fault-tolerance techniques (e.g., ECC, algorithmic redundancy) with FHE-specific strategies (e.g., ciphertext integrity checks, noise-aware recovery). Contribution/Results: Our results reveal that FHE exhibits high and scheme-dependent sensitivity to bit flips; standard hardware- or software-level mechanisms fail to balance efficiency and correctness. In contrast, noise-aware fault-tolerant designs significantly improve decryption success rates without prohibitive overhead. This work provides empirical foundations and concrete design guidelines for building dependable FHE systems in real-world deployments.

Technology Category

• Machine Learning: Hardware-aware ML
• Computer Vision: Adversarial Attacks & Robustness
• Natural Language Processing: Safety and Robustness

Application Category

• Security and Privacy: Large-scale security measurements
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Fully Homomorphic Encryption (FHE) represents a paradigm shift in cryptography, enabling computation directly on encrypted data and unlocking privacy-critical computation. Despite being increasingly deployed in real platforms, the reliability aspects of FHE systems, especially how they respond to faults, have been mostly neglected. This paper aims to better understand of how FHE computation behaves in the presence of memory faults, both in terms of individual operations as well as at the level of applications, for different FHE schemes. Finally, we investigate how effective traditional and FHE-specific fault mitigation techniques are.

Problem

• Analyzing FHE system reliability under memory faults
• Evaluating fault impact on operations and applications
• Assessing effectiveness of fault mitigation techniques

Innovation

• Analyzes FHE system behavior under memory faults
• Evaluates fault impact across operations and applications
• Tests traditional and FHE-specific mitigation techniques