DynaFog: A Dynamic Task Offloading Framework for IoT-Based Fog Computing Platforms

This paper explores the complexities of task offloading in a dynamic fog computing network, where multihop Internet of Things (IoT) access points (APs) support the connectivity between IoT devices and fog computing nodes. The suggested DynaFog framework is organised around key elements in the context of IoT architectures based on fog computing. These components include the selection of fog nodes with skill, the optimization of paths for effective offloading, and the best decision-making about local or remote task computation. We carefully formulated the task offloading problem over a multi-hop communication link by considering an integer linear program (ILP) in order to handle this intricate scenario. Owing to the non-convexity of the possible solution set, a greedy approach-driven heuristic strategy has been proposed to efficiently navigate the problem space. The technique for the solution incorporates a holistic viewpoint that covers multi-hop routing, energy utilization, delay, and dynamic network characteristics, particularly with regard to utilization of link and fog controllers' capacity. Our experimental results validate the effectiveness of the suggested strategy, demonstrating a noteworthy decrease in average delay and energy usage in comparison to benchmarks.