Containers storage optimization of lightweight CaaS for embedded systems
No Thumbnail Available
URL
Journal Title
Journal ISSN
Volume Title
School of Electrical Engineering |
Master's thesis
Authors
Date
2024-12-18
Department
Major/Subject
Cloud and Network Infrastructures
Mcode
Degree programme
Master's Programme in ICT Innovation
Language
en
Pages
72
Series
Abstract
The Internet of Things (IoT) and Multi-Access Edge Computing (MEC) are pivotal technologies in the evolution of 5G networks, enabling real-time and low-latency services. With IoT rapid expansion came unprecedented data growth, creating the need for scalable and resource-efficient solutions. As a solution, MEC deploys computational power closer to the data source, using embedded systems, such as Base Transceiver Stations (BTS), to support edge computing. However, the constrained resources of BTS hardware highlight the need for lightweight container orchestration platforms. The topic of this thesis is driven by this growing importance of lightweight Container as a Service (CaaS) in resource-constrained environments. With lightweight Kubernetes distributions, such as K3s, a reduced resource usage as well as more flexibility for customization are offered making them theoretically more suitable for Edge based environments. The objective of this study is to address the suitability of K3s for BTS and to propose an optimization solution for limited storage which can block the deployment of services. The study approaches the problem by first analyzing the evolution of mobile networks and MEC architecture to define the purpose behind the limitations of BTS hardware. Afterwards, research on container workflows to identify opportunities for optimization is led to establish a course of action in the implementation of a design aiming to optimize storage usage. The results of the thesis display an improved resource usage for the use of snapshots instead of container image, with however, a higher latency. With the usage of compression algorithm, the latency is greatly reduced and the resource usage worsen. Overall, the results show a margin of improvement in the current container workflow.Description
Supervisor
Manner, JukkaThesis advisor
Zizka, JanKeywords
cloud, kubernetes, lightweight kubernetes, edge computing, containerization, mobile network