Background

Proxy servers are widely used for legitimate purposes such as privacy and load balancing. However, a new class of proxies, known as residential proxies, has emerged, which are difficult to detect and increasingly used for malicious purposes such as web scraping, fraud, and cyberattacks. Unlike traditional proxies, residential proxies route traffic through residential IP addresses, making them appear as legitimate users and posing significant challenges to detection systems.

Project scope

The goal of this thesis is to evaluate and improve an existing system to detect residential proxies.

This includes for example working the following tasks:

• Setting up a global evaluation environment using cloud providers
• Evaluating and improving the existing system (e.g., to make it less susceptible to changing network environments)
• Benchmarking the improved system against other existing tools

Required skills

• Detailed knowledge in networking protocols (in particular IP, TCP, HTTP)
• Basic knowledge in machine learning (e.g., clustering algorithms)

Availability

This project is available for a start in February 2025 or later.