DeAI Research
Source · Journal of Communications Software and Systems

HW/SW Architecture Exploration for an Efficient Implementation of the Secure Hash Algorithm SHA-256

2021Manel Kammoun; Manel Elleuchi; Mohamed Abid; Abdulfattah M. Obeid

cid a101b1578faf78d948d8749cbad58ac0401aeb515e9d140dc32353dbcf0558ea · original · open access · via openalex

Title: HW/SW Architecture Exploration for an Efficient Implementation of the Secure Hash Algorithm SHA-256
Authors: Manel Kammoun; Manel Elleuchi; Mohamed Abid; Abdulfattah M. Obeid
Venue: Journal of Communications Software and Systems
Year: 2021
Canonical: doi:10.24138/jcomss-2021-0006
URL: https://doi.org/10.24138/jcomss-2021-0006
Collector: openalex
OA: true

ABSTRACT
Hash functions are used in the majority of security protocol to guarantee the integrity and the authenticity. Among the most important hash functions is the SHA-2 family, which offers higher security and solved the insecurity problems of other popular algorithms as MD5, SHA-1 and SHA-0. However, theses security algorithms are characterized by a certain amount of complex computations and consume a lot of energy. In order to reduce the power consumption as required in the majority of embedded applications, a solution consists to exploit a critical part on accelerator (hardware). In this paper, we propose a hardware/software exploration for the implementation of SHA256 algorithm. For hardware design, two principal design methods are proceeded: Low level synthesis (LLS) and high level synthesis (HLS). The exploration allows the evaluation of performances in term of area, throughput and power consumption. The synthesis results under Zynq 7000 based-FPGA reflect a significant improvement of about 80% and 15% respectively in FPGA resources and throughput for the LLS hardware design compared to HLS solution. For better efficiency, hardware IPs are deduced and implemented within HW/SW system on chip. The experiments are performed using Xilinx ZC 702-based platform. The HW/SW LLS design records a gain of 10% to 25% in term of execution time and 73% in term of power consumption.