Network Intrusion Detection System for Denial-of-Service attack detection in 5G
The number of connected devices in the network continues to increase year by year, specially with the introduction of fifth-generation (5G) technology, which offers higher capacity to accommodate the growing demand. However, these devices, such as Internet of Things (IoT), are vulnerable to Denial o...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2024 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/417190 |
| Acceso en línea: | https://hdl.handle.net/2117/417190 |
| Access Level: | acceso abierto |
| Palabra clave: | Machine learning Deep learning (Machine learning) Computer security Sistema de Detecció d'Intrusions Aprenentatge Federat 5G Aprenentatge Automàtic Aprenentatge Profund DoS DDoS Denegació de Servei Ciberseguretat Xarxa Intrusion Detection System Federated Learning Machine Learning Deep Learning Denial of Service Cybersecurity Network Aprenentatge automàtic Aprenentatge profund Seguretat informàtica Àrees temàtiques de la UPC::Informàtica::Seguretat informàtica Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàtic |
| Sumario: | The number of connected devices in the network continues to increase year by year, specially with the introduction of fifth-generation (5G) technology, which offers higher capacity to accommodate the growing demand. However, these devices, such as Internet of Things (IoT), are vulnerable to Denial of Service (DoS) attacks if they are not properly secured. A DoS attack inundates a network or device with excessive traffic, overwhelming it until it is inaccessible to legitimate users. This vulnerability poses significant risks to critical services such as healthcare, energy, and transportation. This thesis addresses the challenge of detecting DoS attacks in 5G networks by designing and developing an Intrusion Detection System (IDS) based on Deep Learning (DL). The proposed IDS monitors network traffic in real-time to identify DoS patterns and alerts administrators when potential attacks are detected. The designed IDS is composed of two neural network models: the ADC model, which classifies benign network traffic flows from flows containing DoS patterns, and the DoSC model, which categorizes the specific type of DoS attack. These ML models are trained using the Federated Learning (FL) paradigm, which involves three clients, each utilizing a portion of data from a public 5G network traffic dataset. This approach enables the IDS to learn from diverse data sources without compromising data privacy. The IDS models were evaluated on unseen data, achieving an accuracy of 100%, which demonstrates the high capability of the IDS to detect DoS patterns in network flows. The developed IDS was deployed in a simulated environment, where it receives network traffic flows, analyses the network to alert administrators upon detection, and provides an interface for monitoring the detected DoS attack flows. |
|---|