CBL - Campus del Baix Llobregat

Projecte matriculat

Títol: Machine-learning-based IP traffic detection


Departament: ENTEL

Títol: Machine-learning-based IP traffic detection

Data inici oferta: 04-02-2024     Data finalització oferta: 04-10-2024

Estudis d'assignació del projecte:
    MU MASTEAM 2015
Tipus: Individual
Lloc de realització: EETAC
Paraules clau:
Internet traffic, Machine Learning, firewalling, detection, network security
Descripció del contingut i pla d'activitats:
Detecting and identifying IP traffic through firewalls and routers is crucial for maintaining network security and ensuring proper network management. By analyzing incoming and outgoing traffic, these devices can filter out malicious packets, prevent unauthorized access, monitor network usage patterns and manage network resources accordingly. This capability safeguards sensitive data, protects against cyberattacks, and optimizes network performance.

Typically, Internet flow detection follows static rules based on reading certain header fields (e.g. TCP/UDP ports and flags). However, traffic is constantly changing, as new devices are added to networks and new applications are used. This can make it difficult to keep up with the latest trends and to create rules that will still be effective over time. Besides, this fields could be forged to hide malicious traffic.

This project explores the application of machine learning techniques for the detection and identification of distinct traffic flows generated by various Internet applications. In the initial phase, diverse traffic traces will be generated using different applications, operating under varying channel conditions. The collection and analysis of these traffic traces will be conducted using tools such as Wireshark, while netem will be employed to synthetically recreate different network conditions. The acquired data will then undergo processing to extract relevant features that will serve as inputs for training various machine learning models. The overarching objective is to develop a machine learning model proficient in accurately identifying Internet applications through traffic inspection.

Orientació a l'estudiant:
The student will possess a strong foundation in machine learning and proficiency in Python programming. Additionally, familiarity with network analysis tools like Wireshark and netem would be considered advantageous.
Horari d'atenció a estudiants per a l'assignació de projecte:

© CBLTIC Campus del Baix Llobregat - UPC