Deep Learning-Based Offload and Load Balancing in IoT Fog

EasyChair Preprint 14851

Abstract

In this paper, we present our doctoral research that addresses the problem of resource management in IoT-Fog platforms. Several approaches are proposed, including task offloading, resource allocation, etc., our approach will be based on resource prediction. We hope that this approach will help IOT devices improve their behavior in finding available resources either locally or in the Fog tier. Collecting a dataset of the processing history of jobs in the IOT-Fog environment will allow us to predict the load level of the devices and then the resources needed to process these workloads using one of the deep learning tools such as LSTM for example. The combination of this tool with one of the already existing approaches such as Deep Reinforcement Learning will allow us to automate the management of resources in the IoT-Fog environment in a more suitable way.

Keyphrases: Fog Computing, Internet of Things (IoT)., deep learning, resources prediction.

@booklet{EasyChair:14851,
  author    = {Abdellahi Krama and Sofiane Ouni},
  title     = {Deep Learning-Based Offload and Load Balancing in IoT Fog},
  howpublished = {EasyChair Preprint 14851},
  year      = {EasyChair, 2024}}