Performance Evaluation of Response Based Cryptography Versus Fuzzy Extractors Based on Error Correction Codes

In today’s interconnected world, securing systems stands as a critical imperative. Zero-Knowledge Proof (ZKP) is a pivotal cryptographic method that validates information authenticity without revealing supplementary data that goes beyond the validity of the claim. When seamlessly integrated with Physical Unclonable Functions (PUFs), a robust Zero-Knowledge Trust Model materializes, elevating both security and trust within systems. It’s crucial to note that responses derived from PUFs can be susceptible to alterations due to environmental or other external factors, resulting in increased error rates. The traditional method of rectifying errors within responses received after challenging the PUF has been to implement Fuzzy Extractors based on Error Correction Codes. This paper conducts a comparative analysis between Response-Based Cryptography (RBC) and Fuzzy Extractors, assessing their efficacy in terms of latency induced by each protocol when correcting an identical percentage of errors, and evaluates the security of each based on the required entropy/number of bits essential for system security. Our investigation reveals that RBC exhibits robust competitiveness against Fuzzy Extractors across varying noise levels. Specifically, RBC surpasses Fuzzy Extractors at lower noise levels while maintaining high key entropy. Conversely, Fuzzy Extractor methods demonstrate superior error correction efficiency at higher noise levels but compromise system entropy due to reliance on a shortened random number to rectify more bits in the key. Consequently, RBC emerges as a highly secure approach for rectifying PUF-based keys, notably outperforming Fuzzy Extractors at significantly low noise levels with a latency superiority of an order of 10⁻².