HQC Post-Quantum Cryptography Decryption with Soft-Decision Reed-Solomon Decoder

The Need

Post-quantum cryptography must deliver strong security without imposing excessive key sizes, memory requirements, silicon area, or processing latency. In Hamming Quasi-Cyclic (HQC) cryptography, conventional hard-decision error correction requires relatively long codewords to maintain extremely low decryption-failure rates. These longer codewords increase key and ciphertext sizes and contribute to hardware and memory overhead, creating implementation challenges for resource-constrained and high-performance cryptographic systems.

The Technology

This technology, developed by OSU engineers, improves HQC decryption by using reliability information already available during error-correction processing instead of treating every decoded symbol equally. Hardware-efficient soft-decision decoding uses this information to improve error-correction performance, allowing HQC to meet its required decryption-failure target with shorter codewords. The latest demonstrated approach applies generalized minimum-distance decoding optimized for HQC's code structure, reducing overall key length while maintaining the required security-related decoding performance.

Commercial Applications

  • Hardware accelerators for post-quantum cryptography
  • Security processors, cryptographic coprocessors, and secure semiconductor IP
  • Resource-constrained embedded and edge devices implementing HQC
  • Hardware security platforms requiring efficient quantum-resistant key establishment

Benefits/Advantages

  • Smaller keys: For the evaluated HQC-128 design, key length was reduced by 22% compared with hard-decision decoding
  • Lower implementation area: Synthesized HQC-128 results showed a 15% reduction in overall area compared with evaluated hard-decision implementations
  • Faster decryption: The same comparison demonstrated approximately 20% lower latency
  • Efficient use of existing information: The approach exploits reliability information produced during decoding to improve error correction without simply increasing codeword length

Loading icon