A Cross-Environment Evaluation Framework for Container Orchestration Tools: Optimizing Efficiency and Implementation Trade-offs for Real-World Workloads

Ali, Mohd Sajid (2025) A Cross-Environment Evaluation Framework for Container Orchestration Tools: Optimizing Efficiency and Implementation Trade-offs for Real-World Workloads. Masters thesis, Dublin, National College of Ireland.

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

Abstract

The choice of a container orchestration tool, whether in an on-premises environment, cloud environment, or a hybrid environment. It is difficult due to the fact that performance, resilience, and the operational effort heavily rely on the platform choice, as well as the deployment model. In this project, four orchestrators, which are Kubernetes (k3s), Docker Swarm, HashiCorp Nomad, and Red Hat MicroShift, will be empirically studied in a realistic environment. It uses a realistic multi-container Django Expense Tracker Application with PostgreSQL under synchronous HTTP CRUD workloads. The stack was tested on on-prem VMs, on the AWS EC2, and Hybrid setup, respectively, with JMeter baseline, induced-fault, and spike scenarios. The metrics measured through Prometheus and automation scripts are HTTP latency, throughput, error rates, CPU and memory consumption, HTTP up-time, deployment time, autoscaling lag, failure-recovery time, and storage overhead. The Docker Swarm had the lowest on-prem baseline latency (~0.7ms versus ~6.2ms with Kubernetes) and deployments (~30s versus ~60s). While MicroShift consistently consumed ~2GB RAM but could maintain a higher reliability. ANOVA with one-way showed statistically significant (p<0.01) differences between lightweight and Kubernetes-based platforms. Rather than suggesting a new orchestrator, this study focuses on comparative evidence to support decision making. Kubernetes/MicroShift are suitable to mission-critical, fluctuating loads, and Swarm/Nomad to smaller and steady-state services where easiness and efficiency are key fronts.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Makki, Ahmed UNSPECIFIED
Uncontrolled Keywords:	Container orchestration; Kubernetes; Docker Swarm; HashiCorp Nomad; MicroShift; hybrid cloud; performance evaluation; autoscaling
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
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Ciara O'Brien
Date Deposited:	31 Aug 2026 11:55
Last Modified:	31 Aug 2026 11:55
URI:	https://norma.ncirl.ie/id/eprint/9687

Actions (login required)

View Item