Evaluating the Quality of Autonomic Internet of Things Applications

The rapid evolution of the Internet of Things (IoT) has facilitated the development and deployment of IoT applications in domains such as manufacturing, smart cities, retail, agriculture, health etc. Such IoT applications often are collecting data, responsible for analyzing, and for extracting insightful information to enable decision-making and actuation. In such IoT environments with minimal/no human intervention, evaluating quality is a grand challenge. The notion of quality is essential in order to ensure that autonomic IoT applications are resilient to distributed, uncertain, and heterogeneous IoT environments. The current approaches of evaluating quality in the literature, such as multimedia IoT, involve human feedback (i.e., Likert scale, MOS, etc.), which may be unsuitable for autonomic IoT applications with minimal/no human intervention. In this thesis, we have introduced and proposed a pioneering definition for autonomic IoT applications. We also introduce a novel definition for quality of autonomic IoT applications. We have proposed, implemented, and evaluated a novel technique for the holistic evaluation of the quality of autonomic IoT applications. The proposed solution is based on first identifying the key metrics which we name IoT quality metrics that contribute to the quality of autonomic IoT application life cycle (data sensing, sensed data transmission, data analytics, analyzed information transmission, decision-making/actuation). Finally, we propose, implement, and evaluate a novel framework, namely Internet of Things-Quality Watch (IoT-QWatch), that supports the development of quality aware autonomic IoT applications.