Analyzing the Resource Utilization of AES Encryption on IoT Devices

With the prosperity of the Internet of Things (IoT), the implied security issues demand heightened attention. However, conventional cryptographic security solutions often lead to heavy computations and high energy costs, which can hardly be afforded by resource-constrained IoT edge devices. Therefore, it is essential to retrieve real testing data and investigate the tradeoff between information security and its resource consumption on IoT edge devices. This paper explores the duration and energy consumption of the Advanced Encryption Standard (AES), implemented through both software and hardware with various key and buffer size settings on two resource-constrained IoT edge devices. In particular, we observe that (1) compared to software, the hardware implementation is more sensitive to buffer size settings and only consumes lower overall time and energy when the buffer size is sufficiently large; (2) the security premium provided by an increase in key size leads to increased resource consumption in all cases; and (3) comparing the two IoT boards, the CYW board, which is designed with faster default CPU clock rate and more memory, consumes fewer resources overall than the BCM board. These observations not only help advance the understanding of the trade-off between IoT devices' security needs and resource consumption, but also shed light on future improvement of light-weight security designs.