Dynamic Multi-Hash Cryptographic Engine on Zynq FPGA for Energy-Conscious Edge AI Applications

Abstract

With the increasing demand for energy-efficient edge computing, maintaining secure data processing without sacrificing performance has become a critical requirement. The proposed system introduces a flexible cryptographic engine implemented on the Zynq-7000 FPGA, capable of supporting multiple hashing algorithms, including SHA-2, SHA-3, SHAKE, and BLAKE2, within a unified hardware architecture. The proposed system introduces runtime adaptability, allowing dynamic selection of the most suitable hashing algorithm based on current workload conditions. This eliminates the need for hardware reconfiguration and enhances system flexibility in real-time applications. The design incorporates energy-aware techniques such as clock gating and partial reconfiguration, while a control-driven finite state mechanism manages transitions between operating modes to achieve an effective balance between performance and power consumption. Experimental validation confirms that the system achieves throughput up to 1.25 Gbps while maintaining power consumption below 1 W. Under low-power operating conditions, the design demonstrates energy savings exceeding 40 % compared to conventional static implementations. The proposed architecture provides a scalable and sustainable solution for secure data processing in edge computing environments, particularly in applications requiring real-time adaptability and efficient resource utilisation.