Thesis (Ph.D.) - Cairo University - Faculty of Computers and Artificial Intelligence - Department of Operations Research and Decision Support

The research of this Ph.D. is fulfilled in the context of the optimized monitoring of Internet of Things (IoT) networks. The IoT enables the interconnection of billions of sensors, actuators, even humans to the Internet, creating a wide range of services, some of which are mission-critical. However, IoT networks are faulty in nature; Things are resource-constrained in terms of energy and computational capabilities. Moreover, they are connected via lossy links. For IoT systems performing a critical mission, it is crucial to ensure connectivity, availability, and network reliability, which requires proactive network monitoring. The idea is to oversee the network state and functioning of the nodes and links; to ensure the early detection of faults and decrease in node unreachability times. It is imperative to minimize the resulting monitoring energy consumption to allow the IoT network to perform its primary function. Furthermore, to realize the integration of the monitoring mechanism with IoT services, the proposed models should work in tandem with the IoT standardized protocols, especially the IPv6 for Low-power Wireless Personal Area Networks (6LoWPAN) and the Routing Protocol for Low-power and lossy networks (RPL). In this challenging context, the first step of analysis is to ensure the (optimal) placement of monitoring nodes (monitors) to cover the given domain. Leveraging the graph built by RPL (the DODAG), the monitoring coverage can be modeled as the classic Minimum Vertex Cover (MVC) on the DODAG. MVC is NP-hard on general graphs and polynomial-time solvable on trees

Issued also as CD