A scheduling-based dynamic fog computing framework for augmenting resource utilization

Fog computing is one of the most important emerging paradigms in recent technological development. It alleviates several limitations of cloud computing by bringing computation, communication, storage, and real-time services near to the end-users. However, with the rapid development of automation in smart cities, the number of task executions by fog nodes are increasing, requiring additional fog nodes. In this paper, we present a Scheduling-based Dynamic Fog Computing (SDFC) Framework to augment the utilization of existing resources rather than adding further fog resources. It includes an additional layer, Master Fog (MF), between the cloud and general-purpose fogs, which are addressed here as Citizen Fog (CF). The MF is responsible for deciding task execution in CFs and the cloud. We use the Comparative Attributes Algorithm (CAA) to schedule tasks based on their priority and a Linear Attribute Summarized Algorithm (LASA) to select the most available CF with the highest computational ability. Our empirical results validate our SDFC framework and show the dependency on the cloud reduces by 15%–20% and overall execution time decreases by 45%–50%.