An Initial Approach to a Multi-access Edge Computing Reference Architecture Implementation Using Kubernetes

Abstract

The increasing demand of data and real-time analysis has given rise to edge computing, providing benefits such as low latency, efficient bandwidth usage, fine-grained location tracking, or task offloading. Edge computing based on containers brings additional benefits, facilitating the development and deployment of scalable applications adapting to changing market demands. But in order to enable edge computing in the telco industry, it is important that current standardization efforts are followed by software platforms implementing those standards. This paper proposes an approach to the design and implementation of an edge computing platform based on Kubernetes and Helm providing functional blocks and APIs as defined by ETSI in the Multi-Access Edge Computing (MEC) reference architecture. Although this proposal is still at a work-in-progress state, this paper describes the design and implementation of an open-source proof-of-concept scenario focusing on the lifecycle management of cloud native MEC applications. The resulting prototype shows the feasibility of this approach, that can be adequate to create a lightweight MEC demonstration platform for university laboratories and experimentation.