Post-kwantum cryptografie: NIST-standaardisatie en verder

The security of most widely used public-key cryptography is fundamentally based on two computationally intractable problems: the integer factorization problem and the elliptic-curve discrete logarithm problem. However, these problems can be solved easily in polynomial time with the help of large-scale quantum computers using Shor's, Proos and Zalka's algorithm. Post-quantum cryptography provides schemes based on hard problems that are hard-to-solve not only by adversaries with classical computers but also by adversaries with large-scale quantum computers. Lattice-based cryptography is a branch of post-quantum cryptography consisting of schemes based on hard lattice problems. The National Institute of Standards and Technology (NIST) recently published the first set of standard post-quantum schemes. Kyber and HQC are the standard key-encapsulation mechanism/public-key encryption schemes, whereas Dilithium, Sphincs+, and Falcon are the standard digital signature algorithms. Although NIST's selection of the first set of post-quantum cryptographic standards is one of the most significant steps toward the transition from classical public-key cryptography to post-quantum cryptography, the field of post-quantum cryptography is still in a nascent stage compared to the decade-old classical public-key cryptography, such as RSA or ECDSA. Therefore, a major focus of this thesis is to bridge the gap between the theory and practice of post-quantum cryptography by exploring various problems that arise during the real-world deployment of post-quantum cryptography. Specifically, I focus on improving the practicality of lattice-based cryptography by proposing new efficient schemes, designing lightweight post-quantum schemes suitable for resource-constrained devices, efficient implementations, analyzing practical physical attacks on implementations of the post-quantum schemes, and designing their countermeasures to thwart the attacks. The first area of research of this dissertation is the exploration of various design choices for lattice-based key-encapsulation mechanisms to make them suitable in different environments under various operational constraints such as memory, silicon area, execution time, power requirement, etc. In our first contribution, we have explored various design choices for lattice-based key-encapsulation mechanisms, focusing on improving execution time on various hardware and software platforms. Our first design, Scabbard, is a suite of three key-encapsulation mechanisms that improved the state-of-the-art post-quantum key-encapsulation mechanism designs by reducing execution time, stack memory requirements, and/or area footprint. In our second contribution, we performed another design space exploration based on hardware resources (such as block RAM, distributed RAM, look-up tables, etc.) estimation and proposed the lattice-based key-encapsulation mechanism, Rudraksh, for resource-constrained devices. We also substantiate our claim of its usefulness with real-world implementations on an FPGA. The second area of research of this dissertation is the analysis of physical attacks. This attack analysis is essential when the attacker can access a device that executes cryptographic implementations, and such scenarios occur frequently in real-world deployments. Therefore, this is a crucial problem to address before deploying post-quantum cryptographic schemes in real-world applications, and countermeasures against physical attacks need to be developed as a prevention measure. In our third contribution, we proposed side-channel attack-resistant masked implementations of lattice-based KEM, Saber KEM, one of the NIST finalist KEMs. We showed that masked Saber outperforms the state-of-the-art masked implementations of Kyber, one of the NIST standards. Our investigations show that this is largely due to power-of-two moduli. Our final contribution to this thesis is physical attack analyses on implementations of the standard key-encapsulation mechanism Kyber. We proposed a cross-attack, fault-injection attack on the masked implementation of Kyber. We showed that the exploited vulnerability was introduced due to the algorithmic change caused by masking.