Comparative Evaluation of Energy and Water Efficiency Across Cloud Deployment Models

Suresh, Yashashwini (2025) Comparative Evaluation of Energy and Water Efficiency Across Cloud Deployment Models. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (859kB) | Preview

Abstract

The rapid growth of cloud computing has increased the energy and water consumption of the globe but little research has been done comparing the feasibility of the trending deployments models to the same load. This gap is bridged in this paper, which compares environmental performance of three trending compute models, i.e., virtual machine (EC2), containerized services (Docker on EC2) and serverless functions (AWS Lambda). It is to understand how different architectures can impact the sustainability and also ascertain the best efficient model in delivering energy and water reductions at the changing workload conditions. To generate a microservice workload, this research has developed a controlled microservice benchmark to generate the homogenous compute and I/O bounded workloads. The deployment became repeated multiple times and Locust and AWS CloudWatch had been deployed to collect the metrics. Energy usage was estimated with the help of CPU-based power models (EC2 and Docker) and GB-seconds metrics (Lambda), and water usage was estimated with the help of standard datacentre WUE factors. It is found that serverless functions use much less energy and water, significantly, which is several folds less than the amount of the two consumed by traditional servers, particularly at low and medium loads. Containers also outperform low-load virtual machines and scale to high load as hardware usage is which is not markedly better. These findings can be used to reinforce the existing literature on the overheads of virtualization and to highlight that serverless computing is a highly sustainable solution to idle, short-lived workloads. More work should be added in terms of the number of providers, work types, and functions, which take longer periods.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Siddig, Abubakr UNSPECIFIED
Uncontrolled Keywords:	Virtual Machine; Containers; Energy metrics; Water metrics; Serverless Models
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Cloud computing H Social Sciences > HC Economic History and Conditions > Natural resources > Power resources > Energy consumption
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Ciara O'Brien
Date Deposited:	01 Sep 2026 11:35
Last Modified:	01 Sep 2026 11:35
URI:	https://norma.ncirl.ie/id/eprint/9741

Actions (login required)

View Item