Custom Kubernetes Scheduler on Raspberry Pi for IoT Applications, Optimizing Node Selection Based on Energy Efficiency

Venkata Durga Abhinav, Kotagiri (2025) Custom Kubernetes Scheduler on Raspberry Pi for IoT Applications, Optimizing Node Selection Based on Energy Efficiency. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (941kB) | Preview

Abstract

This study proposes the deployment of a Kubernetes cluster using Raspberry Pi devices with a custom Kubernetes scheduler specifically designed for edge environments, with the help of Kubernetes scheduler plugins. The primary objective of the study is to select the nodes by incorporating real-time bandwidth and energy metrics to deploy the workloads, which helps in improving the performance and reducing resource wastage. With today’s growing adoption of the Internet of Things, fog and edge paradigms, there is a demand for efficient and low-power edge computing infrastructure. Raspberry Pi devices, even though they are resource-constrained, offer a cost-effective solution for deploying containerized applications or workloads in an edge environment. However, the default Kubernetes scheduler considers only the CPU and memory utilization metrics, which are not enough for the better optimization of these devices to run IoT applications. In this paper, the methodology involves collecting the custom node metrics using open-source tools such as Prometheus, for monitoring the node’s energy usage, and Iperf for the node’s bandwidth measurement, these metrics are used and integrated with a plugin-based scheduler framework to score and rank nodes more intelligently than the default Kubernetes scheduler. Through the experiment, by comparing the default scheduler with the new scheduler, the new scheduler shows measurable improvements in bandwidth cost efficiency and energy-aware workload distribution in the Raspberry Pis. Results are visualized with the help of Grafana dashboards, and these findings highlight the effectiveness of the multi-metric scheduling approach compared to the traditional approach in optimizing Kubernetes for fog and edge computing scenarios.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Gupta, Shaguna UNSPECIFIED
Additional Information:	Kubernetes; Raspberry Pi; Bandwidth Optimization; Energy Efficiency; Prometheus; Grafana; iPerf; Container Orchestration
Subjects:	T Technology > T Technology (General) > Information Technology > Cloud computing Q Science > QA Mathematics > Electronic computers. Computer science > Computer Systems > Computers > Electronic data processing > Electronic data processing--Distributed processing > Edge computing T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science > Computer Systems > Computers > Electronic data processing > Electronic data processing--Distributed processing > Edge computing T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Telecommunications > Computer networks > Internet of things
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Ciara O'Brien
Date Deposited:	31 Mar 2026 10:54
Last Modified:	31 Mar 2026 10:54
URI:	https://norma.ncirl.ie/id/eprint/9275

Actions (login required)

View Item