An Efficient Multi-Objective Task Scheduling in Fog-Cloud Environment

The high communication bandwidth between end-devices and cloud servers along with the associated network latency, renders it infeasible for the modern day Internet of Things (IoT) applications to meet their stringent latency requirements. Hence, it is crucial to process application data closer to the network edge, both optimally and efficiently. To address this concern, we first design a multilayer fog-based loT system architecture. Next, we propose an optimal task scheduling scheme to uniformly distribute tasks amongst the fog nodes on the basis of Quality of Service (QoS) requirements of the application. Our proposed scheduling scheme optimizes delay, energy consumption and fog node utilization simultaneously while considering dependencies amongst sub-tasks in a way, such that delay and energy consumption is minimized whereas fog node utilization is maximized. We evaluate the performance of our proposed scheme both theoretically and experimentally through simulation using the iFogSim simulator. Our results establish that the proposed scheduling scheme efficiently distributes tasks amongst the available fog nodes as well as reduces the energy consumption of the fog nodes by ~ 1.26% compared to a state-of-the-art competitor at the cost of a little extra latency.