Abstract
The threat of quantum computers is real and will require significant resources and time for classical systems and applications to prepare for the threat. At the algorithm level, the two most popular public key cryptosystems RSA and ECC are vulnerable to quantum cryptanalysis using Shor’s algorithm, while symmetric key and hash-based systems are weakened by Grover’s algorithm. Less is understood at the application level, if the replacement post-quantum resistant algorithms being evaluated by NIST will be suitable drop-in replacements for RSA and ECC, and what gaps or implementation issues still exist. This thesis takes the quantum readiness of applications and services using digital signatures to the next level by addressing several areas in the migration journey for classical systems to become quantum-secure. We identify gaps in preparation by examining the feasibility of using candidate algorithms from NIST’s post-quantum cryptography standardization exercise on existing real-world cryptographic implementations and by carrying out adversarial modelling to uncover broader security threats. We next solve two challenges in the migration process by identifying PASTA as the most appropriate threat modelling method for organizations to prepare for post-quantum migration, and by proposing a quantum-resistant layering technique using multi-party computation with zero-knowledge proof for communicating parties with different migration schedules to achieve seamless compatibility. Finally, we propose an alternative post-quantum authentication implementation for low-resource devices using Learning Parity with Noise and a new Verifiable Delay Function construction that is quantum secure.