<![CDATA[The widespread use of proxy servers has introduced challenges in managing and securing internet connections, particularly in detecting non-transparent proxies that obscure the originating IP address. Proxy servers, while beneficial for bandwidth management and anonymity, can be exploited for malicious purposes, such as bypassing geo-restrictions or concealing cyberattacks. This study aims to address the gap in identifying proxy usage by providing an organized review of existing detection techniques and proposing a hybrid server-side detection framework. The objectives of the research include identifying and comparing proxy detection methods, developing a hybrid approach using machine learning, and evaluating its effectiveness in enhancing network security. The methodology involves collecting primary data through controlled environments simulating direct and proxy-based connections. A machine learning model, based on the Apriori algorithm, is employed to analyze network traffic patterns and identify characteristics indicative of proxy usage. Attributes such as IP addresses, port numbers, and round-trip times are used to train the model. The proposed framework is tested for its robustness, accuracy, and speed against existing detection methods. The results demonstrate the feasibility of the hybrid approach in improving the detection of non-transparent proxies, particularly those not easily identifiable using conventional techniques. The findings have significant implications for securing server-side infrastructure, aiding in cyber threat mitigation, and enforcing organizational policies. Future research can expand on this framework by testing it against broader proxy types and integrating real-world data to enhance its reliability and scope. This study contributes to advancing cybersecurity practices by addressing a critical challenge in proxy detection.]]>
Copyrights © 2025
