Abstract

The Internet of Things comprises of a system of devices (or objects) connected to the Internet and interacting with each other to satisfy various tasks or goals. These objects could be sensors, actuators, smart phones, smart appliances, etc. With the ever-increasing demand of IoT in daily life as well as in the industry, and billions of devices being connected over the internet, most IoT applications aim for cost and energy efficiency, scalability, and minimal latency in terms of resource provisioning.
To fulfill these requirements, Cloud Computing might prove beneficial. Cloud Computing provides on demand access to configurable computing resources (servers, memory, network, etc.) in the cloud, which require minimal management by the end user. It comprises of three service models, which are: Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). The cloud IaaS aims at an efficient usage of resources. In the specific case of IoT, these resources are the sensing and actuation capabilities. However, there are still many challenges that the design and implementation of an IoT IaaS faces. Some examples are the heterogeneity of the sensors and actuators, orchestration, provision of bare metal access, and also publication and discovery of the capabilities of IoT devices. This thesis aims at the design and implementation of an architecture for IoT IaaS. First, it lays down a set of requirements essential to the architecture. This is followed by a thorough review of the state of the art. Next, it proposes an architecture for IoT IaaS that utilizes node level virtualization for an efficient usage of IoT capabilities. Functional entities are proposed as well as interfaces relying on RESTful Web services. The interfaces include a low-level interface for homogeneously accessing all the heterogenous capabilities of IoT devices, as well as high level interfaces which allow the IoT cloud users (e.g. PaaS or individual applications) to access these capabilities in an efficient manner. We have implemented a prototype using real-life as well as simulated Temperature sensors & Humidity sensors, and EV3 LEGO Mindstorms robots. The architecture is validated by concrete measurements on the prototype and by extensive simulations.