Evaluation of Application and Platform ContinuousDeployment Strategies and Tools over Edge Clusters

Abstract: The purpose of this thesis was the investigation of using Cloud Native tools in an On-prem Edge Deployment. Edge Computing allows software with explicit demands on latency, data management and reliability to utilize Cloud Computing by moving the compute power to the location of need while leaving the management in the Cloud. The use of Cloud Native tools in an Edge Deployment could allow for an easier transition from Cloud to Edge, while also allowing developers to use the tools they are familiar with. Edge Computing is a relatively new field with great potential and a lack of common used designs. Many of the potential use cases for Edge Computing are currently not very explored, one of them being On-prem Edge. Can the usefulness of Kubernetes and Cloud Native tools be extended to On-prem Edge as well as bride the gap between it and the Cloud. The thesis will follow the Design Science Research methodology since it favours unexplored subjects and situations in computer science through the creation and evaluation of a design. This thesis proposes a design that emphasizes the autonomy of the Edge node by allowing it to be the initiator of interactions. This design has shown promising results through testing of the performance of the design and evaluation of its workflow. While autonomy of the Edge node comes at a resource cost, the benefits greatly makes up for these by simplifying management of the node.