Abstract:
As quantum computing continues to evolve, it poses significant risks to current cryptographic
methods, making the transition to post-quantum cryptography essential.
CRYSTALS-Kyber, officially standardized by NIST in August 2024 as ML-KEM, is
a leading solution designed to resist quantum attacks. This thesis investigates an
enhancement to ML-KEM by replacing its existing Keccak-based hashing function
with Ascon, recently selected by NIST for its lightweight cryptographic properties
known for its efficiency, particularly in environments with limited resources like
embedded systems. The core objective of this research is to evaluate the performance
impact of this change. Testing was carried out on a personal laptop, using Kyber’s
original test cases to measure CPU cycles consumed by key cryptographic operations
both before and after replacing Keccak with Ascon. The results demonstrate that the
integration of Ascon significantly improves computational efficiency while maintaining
the cryptographic integrity and security of ML-KEM. Though this work does not aim
to enhance ML-KEM’s security—which is already ensured by its design—it offers a
justification that the substitution of the hashing function does not negatively impact its
cryptographic integrity. The key contribution of this research lies in making ML-KEM
more suitable for resource-constrained environments, particularly embedded systems,
by improving its efficiency and reducing computational overhead. By exploring the
practical benefits of Ascon’s integration into a post-quantum cryptographic standard,
this thesis contributes to the optimization of secure, quantum-resistant lightweight
algorithm for real-world applications, paving the way for its effective use in embedded
systems and similar platforms.
