Système de Détection d’Intrusion basé sur l’apprentissage fédéré dans le Fog Computing.

Abstract

Fog Computing is an emerging paradigm that extends the capabilities of Cloud Computing by deploying computing and storage resources at the network edge, closer to users and connected devices. However, Fog environments are often distributed, heterogeneous, and dynamic, which makes intrusion detection more complex compared to traditional systems. In this thesis, we propose a federated learning-based approach for detecting intrusions in Fog Computing. Federated learning is a machine learning technique that allows models to be trained collaboratively without sharing raw data among nodes. This approach preserves data privacy while enabling the learning of a global model based on the local information of each Fog node. The obtained results demonstrate the potential of federated learning as a promising solution to enhance the security of Fog environments, thereby paving the way for future research in this field.