Multiple Criteria Decision Making-Based Task Offloading and Scheduling in Fog Environment

Abstract

Fog computing is a three-tier architecture that provides an emerging technology aiming to reduce the delay and energy consumption between IoT (end) devices and the cloud. The fog layer is close to IoT devices; hence the tasks of time-sensitive applications are offloaded from the end devices to the fog nodes. Efficient offloading and scheduling of the tasks (i.e., the order in which tasks are executed at a fog node) jointly minimize waiting and response time. Given a set of fog nodes and a set of tasks, how to select a fog node and how to effectively schedule the tasks to minimize delay is a challenging problem due to heterogeneous nature of the fog environment. To deal with this challenge, we need to jointly offload and schedule the tasks by ranking the fog nodes and the tasks respectively. Although some papers have addressed task offloading and scheduling jointly, none of them have used performance-based ranking. In this paper, we propose a scheme that uses the multilevel Multiple Criteria Decision Making (MCDM) technique for fog node selection during offloading and determining order of task execution in scheduling. The proposed scheme is based on Entropy-based Technique for Order of Preference by Similarity to Ideal Solution (E-TOPSIS), which incorporates delay, energy, and reliability to rank the fog nodes as well as tasks. Through extensive simulations, we show that the proposed scheme outperforms some existing (baseline) algorithms.