Paper 2026/1468

Side-Channel Attacks Revisited - an Optimization Problem Perspective: Bootstrapping and Space Reduction

Erez Tamir, Bar-Ilan University
Osnat Keren, Bar-Ilan University
Itamar Levi, Bar-Ilan University
Abstract

Side-channel analysis (SCA) attacks rely on leakage from a target device. It is common to assume that linear operations implemented by XOR gates produce symmetric leakage and carry negligible side-channel information. In practice, leakage from XOR gates produces complex, non-independent, and time-varying asymmetric behavior. The paper introduces Feature Estimation based Attacks (FEbA) -- a dedicated profiling attack that exploits these asymmetries. The attack is versatile; it was demonstrated to be successful against the sharing and refreshing phases in masking-based implementations by greatly narrowing the guessing key space, with no access to intermediate values. Such attacks have implications for designs such as ASCON, GIBBON, and ACE, where XORs that utilize the key are vulnerable to attacks regardless of the inherent SCA protection levels used in them (e.g., sponge $rate$, the leak-free components for re-keying, and masking order $d$). Experimental results indicate that the entropy of a $32$-bit key can be reduced below $1$ bit using (up to) $20,000$ traces from a standalone XOR without any access to intermediate values, or below $500$ traces with access to intermediate values.

Metadata
Available format(s)
PDF
Category
Secret-key cryptography
Publication info
Preprint.
Keywords
Feature Estimation AttackFEbAMaskingProfilingSCASide Channel Analysis AttacksSharing
Contact author(s)
erez tamir @ biu ac il
osnat keren @ biu ac il
itamar levi @ biu ac il
History
2026-07-22: approved
2026-07-18: received
See all versions
Short URL
https://ia.cr/2026/1468
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2026/1468,
      author = {Erez Tamir and Osnat Keren and Itamar Levi},
      title = {Side-Channel Attacks Revisited - an Optimization Problem Perspective: Bootstrapping and Space Reduction},
      howpublished = {Cryptology {ePrint} Archive, Paper 2026/1468},
      year = {2026},
      url = {https://eprint.iacr.org/2026/1468}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.