Resource Management and Scheduling in Fog Computing Using Deep Learning

Abstract

Fog computing is a novel approach that brings the computational capabilities of cloud computing systems closer to end-users. The common practice in the technological field is to utilize the cloud to provide services that facilitate interactions between users and their files anytime and anywhere. However, several issues can arise, such as increased network pressure during high request volumes, leading to delays in response and service quality. Additionally, the distance between the user and the cloud affects service speed, as shorter distances generally yield faster responses from servers. Cisco has introduced an alternative technology known as fog computing, which operates between the user and the cloud to deliver services more quickly and effectively. In this study, we combined fog computing with the K-nearest neighbor (KNN) algorithm as an initial step for classifying distances, extracting proximity information, and determining file sizes to complete tasks for users with minimal time and effort. In the second phase, we implemented deep learning techniques using Python, along with the SimPy simulator for resource management, which also operates in Python.