Background and Motivation
Edge computing has emerged as a critical paradigm for reducing latency and improving performance by bringing computation closer to data sources. By 2025, analysts predict that 50% of enterprises will have adopted edge computing, up from just 20% in 2024. This rapid adoption has created a demand for efficient orchestration frameworks capable of managing containerized applications across resource-constrained, geographically distributed edge nodes.
Traditional cloud orchestration platforms like Kubernetes, while powerful, often prove heavyweight for edge environments characterized by limited computational resources, intermittent connectivity, and heterogeneous hardware. This has led to the development of lightweight Kubernetes distributions such as K3s, MicroK8s, and k0s, specifically designed for edge deployments. Additionally, specialized edge orchestration frameworks like Oakestra have emerged to address the unique challenges of hierarchical edge computing environments.
Increasing use of heterogeneous networks and edge computing infrastructure makes managing network complexities and resource allocation a significant challenge. The performance characteristics of different orchestration frameworks in constrained edge environments remain poorly understood, particularly regarding their resource utilization patterns, scalability limits, and behavior under varying network conditions. Furthermore, by combining automation with robust orchestration, edge computing will support a growing range of applications that demand low latency, including IoT, augmented reality, and autonomous systems.
Current literature lacks comprehensive performance comparisons that examine the internal architectural differences between orchestration frameworks and their impact on edge deployment scenarios. Understanding these performance bottlenecks and architectural trade-offs is crucial for informing future edge orchestration design decisions and helping practitioners select appropriate frameworks for their specific edge computing requirements.
Expected Outcomes
This thesis aims to conduct a comprehensive performance evaluation of orchestration frameworks in constrained, elastic edge environments. The study will focus on comparing Kubernetes, MicroK8s, K3s, and emerging frameworks like Oakestra across multiple dimensions of performance and resource utilization.
The research will establish a controlled edge computing testbed that simulates realistic resource constraints, including limited CPU, memory, and network bandwidth typical of edge deployments. The experimental setup will include scenarios with varying node counts, application workloads, and network conditions to evaluate framework behavior under different operational stresses. Performance metrics will encompass deployment latency, resource overhead, scaling efficiency, fault tolerance, and network utilization patterns.
A key contribution will be the deep architectural analysis of each framework's components, identifying specific bottlenecks in areas such as control plane overhead, container runtime efficiency, service discovery mechanisms, and inter-node communication protocols. The study will quantify the resource footprint of each framework's core components (API server, scheduler, controller manager, etc.) and analyze how these components behave under resource pressure.
The thesis will evaluate framework performance during elastic scaling scenarios, measuring how quickly and efficiently each system can scale applications up and down in response to demand fluctuations. This includes analyzing the overhead of pod scheduling, image pulling, and service mesh configuration changes during scaling events.
Network-aware performance evaluation will examine how each framework handles intermittent connectivity, network partitions, and varying bandwidth conditions common in edge environments. The study will measure the resilience of each orchestration framework to network disruptions and their ability to maintain service availability during connectivity issues.
Based on the experimental findings, the thesis will provide practical guidance for edge computing practitioners, identifying which frameworks are most suitable for different edge deployment scenarios. The research will also highlight fundamental architectural limitations and bottlenecks that future edge orchestration frameworks should address, contributing to the development of more efficient and scalable edge computing solutions.
Requirements
• Strong understanding of container orchestration concepts • Proficiency in Linux system administration and containerization technologies (Docker, containerd) • Programming skills in Python or Go for developing measurement tools and automation scripts • Knowledge of distributed systems concepts and network protocols • Familiarity with cloud-native technologies and microservices architectures • Experience with infrastructure automation tools (Ansible, Terraform) for testbed setup
Related Work
[1] Giovanni Bartolomeo, Mehdi Yosofie, Simon Bäurle, Oliver Haluszczynski, Nitinder Mohan, and Jörg Ott. 2023. Oakestra: A Lightweight Hierarchical Orchestration Framework for Edge Computing. In 2023 USENIX Annual Technical Conference (USENIX ATC 23). USENIX Association, Boston, MA, 135–148. https://www.usenix.org/conference/atc23/presentation/bartolomeo
