Entropy-Validated Post-Quantum Cryptography for Secure Key Management in Cloud Systems

Abstract

The development of post-quantum cryptography (PQC) is crucial because the advent of quantum computers poses a serious threat to traditional cryptographic methods. Despite the fact that PQC methods (such Kyber, NTRU, and Dilithium) are resistant to quantum attacks, they do not inherently ensure that the cryptographic keys are high-entropy, leaving the systems open to side-channel and brute-force attacks. In order to address entropy-aware validation of PQC produced keys, this research suggests a machine learning method. The quality of one-time key material produced by PQC and conventional algorithms is analyzed using Shannon entropy, a statistical measure for unpredictability. The low-entropy keys are identified and quarantined using a One-Class SVM, ensuring only cryptographically strong keys are used. Five cryptographic schemes are used to empirically evaluate the claims. It was found that Dilithium produces the most secure keys among the schemes, with high validation rates and entropy scores. The combination of machine learning and entropy-driven analysis guarantees QRKG's resilience and a scalable strategy for protecting cryptographic infrastructures from post-quantum attackers. With strict key management criteria, the created method can be expanded to blockchain, cloud, and IOT security systems.