Designing and Deploying Internet of Things Applications in the Industry: An Empirical Investigation

Internet of Things (IoT) aims to bring connectivity to almost every object found in the phys-ical space. It extends connectivity to everyday things, opens up the possibility to monitor, track, connect, and interact with industrial assets more eÿciently. In the industry nowadays, we can see connected sensor networks monitor logistics movements, manufacturing machines, and help organizations improve their eÿciency and reduce costs as well. However, designing and implementing an IoT network today is still a very challenging task. We are witnessing a high level of fragmentation in the IoT landscape and developers regularly complain about the diÿculty to integrate diverse technologies of various objects found in IoT systems, and the lack of clear guidelines and–or practices for developing and deploying safe and eÿcient IoT applications. Therefore, analyzing and understanding issues related to the development and deployment of the Internet of Things is utterly important to allow the industry to utilize its fullest potential. In this thesis, we examine IoT practitioners' discussions on the popular Q&A websites, Stack Overflow and Stack Exchange, to understand the challenges and issues that they face when developing and deploying di˙erent IoT applications. Next, we examine the lack of interoper-ability among technologies developed for IoT and study the challenges that their integration poses and provide guidelines for practitioners interested in connecting IoT networks and de-vices to develop various services and applications. Since security issues are center to the success of this technology, we also investigate di˙erent security threats and challenges across di˙erent layers of the architecture of IoT systems and propose countermeasures. Finally, we conduct a series of experiments to understand the advantages and trade-o˙s of serverful and serverless deployments of IoT applications in order to provide practitioners with evidence-based guidelines and recommendations on such deployments. The results presented in this thesis represent a first important step towards a deep understanding of these very promising technologies. We believe that our recommendations and suggestions will help practitioners and technology builders improve the quality of IoT software and systems. We also hope that our results can help IoT communities and consortia establish standards and guidelines for the development, maintenance, and evolution of IoT software and systems.