Performance Analysis of Cloud Hypervisor using Different Workloads in Virtualization

Abstract

In virtualized cloud computing systems, energy conservation is crucial since it may positively impact various factors, such as decreased operating costs, increased system effectiveness, and environmental preservation. Simultaneously, an energy-efficient task scheduling technique is a potential option for achieving these objectives. Unfortunately, mapping cloud resources to user requests to achieve high performance while reducing cloud resource energy consumption within a user-defined timeframe is a significant difficulty. This paper finds cloud hypervisor power and energy charters with power measurements, namely Amazon AWS, Microsoft Azure, and Google Clouds. We will employ computation- and network-intensive workloads to identify several cloud hypervisor power and energy features and the optimal multi-tenant cloud architecture. Finally, the result depends on the hypervisor's current operation state due to the workloads assigned. (1) Each hypervisor exhibits a different energy and power conception rate according to the workload. (2) On all systems, no single primary hypervisor outperforms the others in terms of current energy consumption, despite their energy efficiency being further matched with various workloads and workload levels. The result obtained from this experiment will be helpful for designers and data centre operators to find out which cloud hypervisor is best for their virtual machine scheduling.