Result: An Enhanced RNS-AES Encryption Scheme with CBC Mode and HMAC for Secure and Authenticated Data Protection

Title:
An Enhanced RNS-AES Encryption Scheme with CBC Mode and HMAC for Secure and Authenticated Data Protection
Contributors:
Department of Cyber Security & Computer Engineering Technology, School of Computing and Information Sciences, C. K. Tedam University of Technology and Applied Sciences, Navrongo, Ghana., Department of computer science - Regentropfen University College (RUC), Department of Computer Science, C. K. Tedam University of Technology and Applied Science, Navrongo, Ghana., Department of Computer Science, School of Computing and Information Sciences, C. K. Tedam University of Technology and Applied Sciences, Ghana.
Source:
Earthline Journal of Mathematical Sciences. :1091-1112
Publisher Information:
CCSD, 2025.
Publication Year:
2025
Original Identifier:
HAL: hal-05359497
Document Type:
Journal article<br />Journal articles
Language:
English
ISSN:
2581-8147
Relation:
info:eu-repo/semantics/altIdentifier/doi/10.34198/ejms.15625.10911112
DOI:
10.34198/ejms.15625.10911112
Accession Number:
edshal.hal.05359497v1
Database:
HAL

Further Information

Modern cryptographic systems in cloud and IoT environments must balance strong security with real-time performance, yet existing methods often require trade-offs that sacrifice speed for security or introduce latency through conservative designs. This paper presents RNS-AES-CBC-HMAC, a hybrid framework that integrates Residue Number System (RNS) arithmetic with AES-256 and HMAC-SHA256 to deliver both performance and robust security. Using a balanced modulus set $\{2^n - 1,\, 2^n,\, 2^n + 1\}$ for constant-time, carry-free arithmetic mitigates side-channel risks, while AES-256 in Cipher Block Chaining (CBC) mode ensures confidentiality and HMAC-SHA256 provides message integrity with minimal overhead. Implemented in Python 3.10 with PyCryptodome 3.18.0 and tested on an AMD Ryzen~5~2500U, the framework achieved encryption/decryption latencies of $55$--$593\,\mu\text{s}$ for 4--15 character payloads, representing 99\% improvement over previous RNS-based hybrids. It scales linearly in time and memory $\mathcal{O}(n)$, consumes only $21\,\text{KB}$, and produces ciphertext entropy of $7.999\,\text{bits/byte}$, surpassing NIST SP~800-22 standards. This dual-layer architecture effectively counters both passive and active threats, making it suitable for low-latency IoT edge devices and high-throughput cloud systems, merging theoretical number systems with practical cryptography for real-world deployment.