Optimization of workflow scheduling in an energy-aware cloud environment

Today, we live in an increasingly dematerialized world. The cloud is the generator of the next wave of technologies, the detonator for all cutting-edge developments like artificial intelligence and blockchain. The cloud is mainly composed of datacenters which consume almost 20% of the total energy footprint of the digital world. The host machines that make up these datacenters are becoming more and more powerful and perform calculations more quickly. Indeed, this increase in computing power is accompanied by increased consumption of electrical energy even in the event of inactivity. In this context, we are talking about a very energy-consuming development in terms of electricity consumption and even in terms of CO2emissions, the internet of which pollutes 1.5 times more than air transport. This exponential consumption of digital makes it very hard to feed with renewable energies. This problem leads us to focus on solutions that optimize the use of resources in the cloud environment in order to schedule workflows especially since there are still today few techniques to limit energy consumption. This article is part of the resolution of this problem of minimizing energy consumption by datacenters in the cloud environment. For that, we try to study a solution based on the technique Dynamic Voltage and Frequency Scaling (DVFS) which allows to moderate the frequency of the CPU thus influencing the energy consumption. The solution based on DVFS is compared with other techniques in order to show its power to achieve the key objective of this article which is the optimization of energy consumption.