Abstract

IoT ecosystem refers to web-enabled smart devices that use embedded processors, sensors, and communication hardware to capture, send, and act on data they obtain from their environments. IoT applications range from healthcare to autonomous vehicles. These Applications are composed of interacting components. Provisioning these applications is rather challenging, especially in hybrid cloud and fog settings.
Cloud computing with its three main services (i.e., Infrastructure-as-a-Service, Platform-as-a-Service, and Software-as-a-Service) still faces some challenges. Through the PaaS, application providers can provision (i.e., develop, deploy, manage, and orchestrate) applications in the cloud. However, the wide-area network used to connect the cloud to the end-users might cause high latency which may not be acceptable. Fog computing can reduce this latency induced by distant clouds by enabling the deployment of some application components at the edge of the network while keeping others in the cloud. Existing PaaS solutions (including IoT PaaS solutions) do not enable provisioning of applications with components spanning cloud and fog.
The main contribution of this thesis is twofold. First, a novel IoT PaaS architecture for hybrid cloud and fog environments is proposed. Second, the general architecture is prototyped and measurements are made to evaluate the feasibility of the architecture. The application components are implemented as Virtual Network Function(VNF) and Cloudify, an open-source NFV orchestration framework is used.
The essential PaaS functional entities, such as deployment engine, orchestrator, migration engine, and publish and discovery engine are identified. A concrete mechanism for publication and discovery for existing cloud and fog nodes (both stationary and mobile fog nodes) has been also proposed. In addition, a set of RESTful interfaces is proposed to enable interaction with components internally and externally. For the evaluation part, we have measured the system end-to-end and migration delay for both centralized and distributed PaaS architecture as well as different placement for application components on cloud and fog nodes.