Towards Hyper-efficient IoT Networks Using Fog Paradigm

Abstract

Fog computing was emerged as a treasured paradigm to improve the efficiency of the typical cloud of things (CoT) architecture of the Internet of Things (IoT) networks. Contrasting to the CoT in which the resource-rich high-performance data centers (DCs) are located far from the energy-constrained terminal nodes (TNs), fog nodes (FNs) in fog-enabled architecture provide computing resources in the proximity of the TNs. Therefore, the TNs consume less energy to offload their generated tasks to the FNs rather than the cloud DCs. Moreover, shortening the distance between the TNs and the FNs results in alleviating the transmission latency for the delay-sensitive tasks generated by the TNs. This is more significant for specific applications, such as smart healthcare, search and rescue, and disaster management, wherein making a prompt decision is vital to save lives. However, Fog-IoT networks still suffer from challenges regarding energy efficiency and provisioning quality of service (QoS) requirements, especially in terms of delay and throughput. The motivation behind this thesis is to tackle the corresponding challenges and improve the performance of the Fog-IoT networks. To this end, novel optimization problems, models, methods, and algorithms are proposed that mainly focus on the energy efficiency improvement and QoS provisioning in Fog-IoT networks. Moreover, due to the importance of the mobility of FNs, the contributions of the thesis encompass improving the performance of Fog-IoT networks with respect to both fixed and mobile FNs.