Optimized Deep Neural Network Based Load Balancing in Fog Computing with Robust Dynamic Scheduling Algorithm

Abstract

The load balancing (LB) in FOG architecture is a formidable task with the availability of minimal resources. This field encounters enormous studies that portray the routing of task between the fog physical devices and cloud nodes. The presence of different types of heterogeneity devices are at helm of difficult scheduling process. To allocate LB based on the available resource we propose a novel Deep Neural Network (DNN) based Tuna swarm strategy based Bacterial Foraging optimization algorithm (TBFO) which employs three stages such as monitoring the fog resources, Classification based on deep learning technique and dynamic scheduler that are optimized. With the dynamic scheduling algorithm this work aims to provide LB in real time application. The first stage is to monitor the resources of server and amassed in fog resource table. The next step is to identify the unerring server and is effectuated with the proposed DNN technique. The last stage is to allocate the process to the selected server with the TBFO. It certifies the robust continuous services in the fog LB. Simulations are conducted and compared the outcomes with the existing works and ensures effective load balancing, make span and resource utilization.