Resource provisioning in fog computing

Abstract

The proliferation of the IoT leads to an ever-growing presence of ubiquitous networked devices. Being located at the network edge, these resources can be exploited to execute IoT applications in a distributed manner. This paradigm, which seamlessly considers technology, infrastructure, and computations from the edge to the cloud, is known as fog computing. In order to enable the efficient exploitation of fog-based computational resources, this thesis introduces fog colonies, which resemble mini data centers consisting of devices at the edge of the network and in the cloud. We model resources and applications in fog colonies, formulate and solve the Fog Service Placement Problem to efficiently distribute IoT applications over fog resources. This thesis addresses the fundamental and critical issues that come along with the adoption of fog computing. The conceptual architecture for fog computing becomes the foundation for the fog computing framework FogFrame – a system able to manage and monitor edge and cloud resources in a fog landscape, and to execute IoT applications. FogFrame enables communication and interaction within the fog as well as provides application management, namely decentralized service placement, deployment, and execution. The thesis shows, through experiments in FogFrame, the performance of different resource provisioning approaches in a real-world fog as well as adaptations to volatile changes in the fog, balancing the workload, and recovering from failures.