# Kubernetes in Action: Exploring the Performance of Kubernetes Distributions in the Cloud

**Authors:** Hossein Aqasizade, Ehsan Ataie, Mostafa Bastam  
**arXiv:** [2403.01429](https://arxiv.org/abs/2403.01429)  
**Date:** 2024-03-03  
**Summary page:** https://www.summarizepaper.com/en/paper/2403.01429v1/

## Key points

- Kubernetes is a leading open-source platform for container orchestration
- Four Kubernetes distributions evaluated: Kubeadm, K3s, MicroK8s, and K0s
- Performance evaluation conducted using OpenFaaS on a cluster of computing nodes on CloudLab
- Experiment aims to assess performance under different virtualization modes (HVM and PV) and container runtimes (Docker and Containerd)
- Metrics measured include request rate, CPU utilization, scaling behavior
- Previous studies have compared lightweight Kubernetes distributions in stress scenarios
- Research has explored software aging issues within Kubernetes infrastructure and evaluated serverless platforms in edge computing environments for IoT applications
- Article aims to provide insights into the performance behavior of various Kubernetes distributions for organizations optimizing container orchestration strategies

## Summary

Kubernetes has emerged as a leading open-source platform for container orchestration, providing organizations with efficient management and deployment of containerized applications at scale. This paper focuses on evaluating the performance of four Kubernetes distributions - Kubeadm, K3s, MicroK8s, and K0s - while running OpenFaaS as a containerized service on a cluster of computing nodes on CloudLab. The experiments conducted aim to assess the performance of two virtualization modes supported by Xen - HVM and PV - as well as two integrated container runtimes - Docker and Containerd - across disk-intensive and CPU-intensive workloads. The research methodology involves setting up the Kubernetes distributions after determining the appropriate underlying Xen mode and container runtime. Various metrics such as request rate, CPU utilization, and scaling behavior are measured to evaluate the performance of each distribution. The experimental setup includes hardware and software configurations along with benchmarks to provide a comprehensive analysis. Previous studies have compared lightweight Kubernetes distributions like MicroK8s, K3s, K0s, and MicroShift in stress scenarios. Additionally, assessments have been made on the performance of different Kubernetes distributions using serverless benchmarks on platforms like OpenFaaS. Research has also explored software aging issues within Kubernetes infrastructure and evaluated open-source serverless platforms in edge computing environments for IoT applications. Overall,<DateTime> article aims to contribute valuable insights into the performance behavior of various Kubernetes distributions under different virtualization modes and container runtimes. These findings can inform decision-making processes for organizations looking to optimize their container orchestration strategies.