Paper 2022/1675
SoK: Assisted Fault Simulation - Existing Challenges and Opportunities Offered by AI
, Radboud University Nijmegen, Radboud University NijmegenAbstract
Fault injection attacks have caused implementations to behave unexpectedly, resulting in a spectacular bypass of security features and even the extraction of cryptographic keys. Clearly, developers want to ensure the robustness of the software against faults and eliminate production weaknesses that could lead to exploitation. Several fault simulators have been released that promise cost-effective evaluations against fault attacks. In this paper, we set out to discover how suitable such tools are, for a developer who wishes to create robust software against fault attacks. We found four open-source fault simulators that employ different techniques to navigate faults, which we objectively compare and discuss their benefits and drawbacks. Unfortunately, none of the four open-source fault simulators employ artificial intelligence (AI) techniques. However, AI was successfully applied to improve the fault simulation of cryptographic algorithms, though none of these tools is open source. We suggest improvements to open-source fault simulators inspired by the AI techniques used by cryptographic fault simulators.
Metadata
- Available format(s)
-
PDF
- Category
- Applications
- Publication info
- Published elsewhere. Artificial Intelligence in Hardware Security (AIHWS) 2023
- Keywords
- Fault simulationOpen-sourceSoftwareAI
- Contact author(s)
-
asmita adhikary @ ru nl
ileana buhan @ ru nl - History
- 2023-04-29: revised
- 2022-12-01: received
- See all versions
- Short URL
- https://ia.cr/2022/1675
- License
-
CC BY
BibTeX
@misc{cryptoeprint:2022/1675,
author = {Asmita Adhikary and Ileana Buhan},
title = {{SoK}: Assisted Fault Simulation - Existing Challenges and Opportunities Offered by {AI}},
howpublished = {Cryptology {ePrint} Archive, Paper 2022/1675},
year = {2022},
url = {https://eprint.iacr.org/2022/1675}
}
