Energy Analysis of Cryptographic Algorithms in Server Environment

Cryptographic algorithms have varying resource requirements depending on their complexity, intended application, and the specific characteristics of the targeted computing environment. As they are computationally heavy, understanding the impact of energy consumption is vital for energy efficiency in computing. In this paper, we investigate energy consumption across various cryptographic algorithms, focusing in particular on LED, Piccolo, PRESENT, GIFT, and AES, and we assess their energy efficiency in a server environment.
The objective is to gain insights into the energy consumption patterns of selected algorithms, also aiming to identify the optimal trade-off between energy consumption and performance. By utilizing a monitoring tool to measure the energy consumed by the algorithms during their execution, we monitor data for each algorithm under different configurations. We systematically collect the energy data across various key sizes and implementation types and varying compiler optimization levels.
Our experiments yield detailed insights into the energy consumption of lightweight ciphers, we observe how energy usage varies with different key sizes and the impact of implementation choices, such as table-based, vperm, and bitslice methods. We also show how compiler optimization levels play a crucial role, offering opportunities for tailored configurations that balance energy efficiency and performance across varying cryptographic scenarios. The collected energy and performance data are a first and crucial step toward a more energy efficient deployment of cryptographic primitives in server environments and datacenters.