Deep Learning and Fog Computing: A Review

Abstract

Fog computing (FC) is a new architecture that aims to reduce network pressures throughout the core network as well as the cloud computing (CC) by bringing resource-intensive functions like computation, analytics, connectivity, also storage, nearest to the clients. In their operations, FC systems can make use of intelligence features to reap the benefits of data that is readily accessible with computing resources to be able to resolve the problem of excessive energy use with power for Internet-of-Things (IoT) apps that require speed. It generates large volumes of data, prompting the creation of a growing number of FC apps and services. Furthermore, Deep Learning (DL), an important field, has made significant progress in a variety of research areas, including robotics, face recognition, neuromorphic computing, decision-making, computer graphics, and speech recognition. Several studies have been suggested to look at how to use DL to solve FC issues. DL has become more common these days to improve FC apps as well as provide fog services such as security, resource management, accuracy, delay, and energy reduction, cost, data processing, and traffic modeling. The current review paper will focus on how to provide an overview of DL functions throughout the FC sector. The DL implementation for FC has evolved into powerful clients with services at the highest level, allowing for deeper analytics and mission answers that are more intelligent.