Reducing Energy and Carbon Cost of Cloud Data Centers through Optimal Virtual Machine Placement algorithms and an Automatic Servers Scale Controller Model

Nowadays, one of the most significant challenges in cloud computing is the massive amount of costs that are being paid by cloud service providers for energy consumption and carbon tax. Accordingly, many efforts, such as increasing the utilization of servers, have been done by cloud providers to reduce these costs. However, many studies have focused only on the reduction of energy consumption on a single data center. Nevertheless, in recent years, there have been introduced lots of novel models that their final goal is to minimize the energy and carbon costs of a cloud provider through geographically distributed cloud data centers. However, one of the most significant defects visible in many models is about keeping up all data centers’ servers in the ready state even when they are in the idle state. In this work, we intend to extend an optimal virtual placement machine (VM) placement algorithm via a useful and straightforward model based on an automated server scale controller mechanism and overall CPU utilization of data centers. Finally, we simulated our proposed model with the CloudSimPlus simulator. The results indicate that our model could significantly reduce the energy and carbon costs compared with the previous models.