Automation of Edge Datacenter Infrastructure

Cloud-native technologies, marked by their inherent flexibility, scalability, and cost-effectiveness, have revolutionized the IT infrastructure landscape of large-scale organizations, especially major telecommunications companies (telcos). Among these technologies, Kubernetes (K8s) has emerged as a leading platform for managing containerized workloads and services. This shift to cloud-native solutions becomes crucial for telcos seeking to efficiently manage growing data traffic and provide a wide array of services. One of the critical obstacles in this digital transformation journey involves the complexity associated with managing bare metal infrastructure, particularly in the context of multi-cluster environments at the network edge. This challenge underscores the importance of the work presented in this thesis. Edge networks, situated closer to end-users and crucial for minimizing latency and quick data processing, often include multiple bare metal nodes interconnected by network switches. The configuration of these switches, however, demands meticulous manual work, such as setting up network interfaces, VLANs, routing protocols, and security settings. This process is not only labor-intensive but also highly susceptible to human error, thereby leading to inefficiencies and potential system vulnerabilities. Addressing these concerns, this thesis presents a novel architecture that leverages the principles of Sylva, an initiative by Linux Foundation Europe, and the extensibility of Kubernetes. \textbf{The architecture seeks to automate the traditionally manual process of network configuration in edge sites with bare metal nodes}, reducing human intervention, increasing accuracy, and significantly enhancing operational efficiency. The proposed architecture pivots around two core elements: the Network Operator and the Actuator Operators. The Network Operator utilizes Custom Resource Definitions (CRDs) to outline the intended network topology and configuration, while the Actuator Operators serve as the communicative conduit between the Network Operator and the physical network devices. The architecture's implementation utilizes Kubebuilder, an open-source Software Development Kit (SDK). A significant component of this implementation is the development of an Actuator Operator compatible with Cisco Network Services Orchestrator (NSO), a model-driven software used to automate the provisioning and configuration of network devices. This operator communicates with NSO using RESTCONF APIs, facilitating real-time execution of configuration modifications and maintaining the alignment between the desired and current states of network configurations. In summary, this thesis addresses a significant challenge in the telecommunications networking domain, providing a cloud-native and Kubernetes-based automated solution for managing and configuring network resources at the edge sites. By resolving the complexities of bare metal nodes' setup, it contributes a meaningful stride towards the digital transformation of network infrastructure management.