Modern Firewall System and AI-Driven Intrusion Detection: Implementation and Evaluation

Abstract

This project is proposed by SOCARAM SPA, a company specialized in various technical services and solutions.It starts by an understanding of the company’s infrastructure and integrating a new őrewall to enhance security. It consists as well of conducting a study on how machine learning can enhance intrusion detection systems by analyzing őrewall logs. We focus on logs from the Sophos XG őrewall, which offer a comprehensive view of network activity. Our goal is to successfully update the infrastructure and identify patterns and anomalies that signal security threats using various machine learning models. We start with an overview of őrewall technologies, discussing their types, methodologies, and policies. We then detail the features and deployment options of the Sophos XG őrewall. Next, we explore key concepts in machine learning and deep learning, emphasizing their relevance to network security. Through our experiments, we evaluate the performance of different machine learning models in detecting intrusions. We assess these models using metrics such as accuracy, precision, recall, and silhouette score. Our results demonstrate the potential of machine learning to enhance classical őrewall systems, making them more effective at identifying and responding to security threats.