Predictive Resource Demand Based Load Balancing Strategy for Fog Computing

Abstract

 The application deployment industry has come a long way in the last decade from Centralized to distributed ,from distributed to cloudand from cloud to fog computing. With the increase in the movable devices and increase in the processing capabilities of these devices, the diversity of the applications has also increased. The recent trends in the application demonstrates a huge distribution in data and processing of the same data through these distributed fog computing nodes. Nonetheless, a typical angle is an obstruction in exact substance transfer, an issue that has been handled with the making of measurements that endeavour to improve accuracy.However, the traditional problems like load balancing still persists on fog processing or fog computing as well. Also, it is to be noted that, the resource distribution is also making the balancing of nodes, much more critical. A good number of recent research attempts have tried to build a generic solution for the load balancing on fog architecture. Nevertheless, the factor of resource distribution has made it highly complex to address to the need. Henceforth, this work aims to deploy a novel strategy using the prediction on the resource demand and further assigning or migrating the balancing tasks to the specific node with the resource availability. The testing of the application has demonstrated higher results compared to the other parallel result outcomes as the proposed method demonstrates SLA performance improvements by 48%, SLA time improvement by 17%, execution time reduction by 63% and number of VM migration reduction by 53%.