DSLHash: Lightweight Hash Function Based on SVSC 4D Chaotic System and Dynamic Substitution Layers

Abstract

Data integrity remains a critical requirement for secure communication in IoT environments. To address this challenge for resource-constrained IoT devices, lightweight cryptographic hash functions have been widely studied. However, static substitution layers in conventional hash functions might be vulnerable to advanced cryptanalysis. Dynamic substitution layers are constructed by exploring several methods such as key dependency, chaotic maps, DNA computations, and elliptic curves. Despite these efforts, existing dynamic substitution layers are not suitable for resource-efficient implementation. To achieve both enhanced security and practicality, this paper proposes a lightweight hash function based on a structure-varying self-coupled (SVSC) chaotic system and a dynamic substitution mechanism. A SVSC 4D chaotic map is first constructed using lightweight arithmetic and logical operations, including modular addition, bitwise shifts, and bitwise negation, thereby avoiding multiplication operations and enabling efficient hardware implementation. The chaotic map is then integrated with a generalized Feistel structure to construct a dynamic substitution layer, where the transformation behavior dynamically changes according to the internal chaotic state. This mechanism enhances nonlinear transformation diversity and contributes to improved diffusion characteristics. Finally, the dynamic substitution layer is incorporated into a sponge-based hash framework supporting flexible digest lengths. The results demonstrate that the proposed SVSC 4D chaotic system exhibits long cycle lengths in a 32-bit finite-precision implementation, mitigating the degradation caused by finite-precision effects and improving its suitability for practical cryptographic applications. Experimental evaluations demonstrate that the proposed methodology achieves a balanced tradeoff among security performance, hardware efficiency, and implementation complexity. Based on a 256-bit sponge state with a 192-bit capacity, DSLHash provides approximately 96-bit collision resistance, which is selected to balance security requirements and hardware resource consumption in lightweight applications.