Load Balancing in Fog Computing Using QoS

Abstract

Various IoT gadgets are ceaselessly increasing day by day and producing an enormous volume of raw data. All of this produced information is passed to cloud servers for processing. As this process adds delay to processing, so it is not suitable for certain applications as some applications require a speedy response. To overcome this condition, fog computing comes in existent, which is an extension to cloud computing. Also, in the present time, fog is the most popular technology due to the vast demand for IoT devices. The fog nodes are placed between IoT devices and the cloud servers. As the execution of the request is performed at the fog layer so it can work with a limited number of resources i.e., less bandwidth, cost, and time as the processing is pushed closer to the end clients. The most challenging task in a fog environment is to appropriately distribute workload among computing nodes during the execution of IoT applications as it is one of the important factors which affect resource efficiency. The performance of any computing paradigm is directly proportional to the load balancing handling mechanism; poor mechanism reduces the overall performance of any computing environment. Realizing the challenge of load balancing among the computing nodes in the fog environment, various mechanisms and methods have been proposed so far and various experiments have also been conducted by the researchers to check the effectiveness of the mechanism. The appropriate load balancing mechanism will increase the effectiveness of the fog system due to better resource utilization. The chapter presents a framework (OLBA) for Load Balancing in Fog computing environments to balance the load between fog devices and improves QoS parameters i.e., Turnaround time resource utilization, response time, and delay parameter. This approach is based on Particle Swarm Optimization (PSO) technique to find the local best and then to compare all the local best to find the ultimate global best solution. An analysis and comparison with the traditional techniques, i.e., FCFS, SJF, Max_Min is also performed for a better understanding of load balancing mechanism in Fog Computing.