Toward the Identification of Anonymous Web Proxies

Anonymous proxies have recently emerged as very effective tools for Internet misuse ranging from Activity and Online Information Abuse to Criminal and Cybersexual Internet Abuse . The ease with which existing proxies can be found and accessed, and new ones can be quickly set up poses an increasing difficulty to identify them. The traditional solution relies on URL filtering approach based on keyword databases. However, such approach cannot keep up with hundreds of new proxies created each day and more importantly the growing adoption of encrypted connections. This work introduces a new methodology that uses flow features to create a server behavior model to identify potential proxies within the observed traffic