Intelligent Latency-Aware Tasks Prioritization and Offloading Strategy in Distributed Fog-Cloud of Things

Abstract

Offloading the dynamic tasks with fog computing is envisioned as a viable option for prolonging resource-limited constraints and improving the computational and communicational latency for delay-sensitive IoT applications. Besides, the priority of tasks and the target layers for offloading them to minimize the incurred service latency is a prime concern in layered computing architecture. To leverage the efficiency of the underlying computing nodes for the tasks’ heterogeneity and computational requirements with deadline constraints, this article presents a fuzzy logic technique to prioritize the tasks based on their resource requirements and associated deadline. For efficient scheduling, an elitism-based multipopulation Jaya is proposed to map these disparate groups of tasks to a cluster amalgamation of computational-rich heterogeneous computing nodes. Moreover, a compatibility-based heuristic offloading strategy is devised to determine compatible computing nodes to offload the computations considering the availability of resources and communicational time from the respective IoT devices. Finally, extensive simulations are carried out with conflicting scheduling parameters appraising the efficacy of the proposed strategy over existing algorithms. The percentages of improvements of the proposed algorithm over the compared algorithms are 35% and 28% for average waiting. time and average service latency, respectively.