Fog Clustering-based Architecture for Load Balancing in Scientific Workflows

Abstract

The rapid growth of intelligent devices accessing cloud data centers leads to network congestion and increased latency. Fog computing provides a ubiquitous and distributed environment in which different fog nodes are deployed near end-users to resolve cloud data centers’ high latency problems, leading to reduced network traffic and latency. Nevertheless, it is a challenging task for fog layer resources to meet complex service quality constraints. Moreover, some objectives are considered to be achieved during the processing of complex and large workflow tasks, i.e., increased energy consumption, less utilization of resources. We have investigated the equal distribution of scientific workflow tasks among available resources to utilize resources and provide energy-aware load balancing properly. In this article, fog computing-based load balancing architecture has been proposed to enhance resource utilization in scientific workflow applications. We proposed a hybrid load balancing algorithm for optimum resource utilization in a fog environment. Our proposed algorithm improves resource utilization and reduces energy consumption as compared to the existing approach. For evaluation of the proposed approach, iFogSim has been used. The article concludes by providing directions for the future researchers.