Detecting Third-Party User Trackers with Cookie Files

nowadays, privacy violations caused by third party services are usually addressed with blacklist approaches. Those approaches have limitation, as they are hard to maintain and need to be frequently updated to provide relevant information. Knowing these limitations, we study ability to identify third-party web trackers having corpus of http request between web resources and associated cookies. Cookies are one of the most popular and easy to use mechanisms to link same user visiting different domains. We propose a method for automatic identification of web-tracking requests, which detects 95% of third-party trackers. The proposed method successfully works with known services and is able to detect previously unseen trackers. To evaluate the proposed technique we artificially collected and then analyzed corpus of internet traffic where we found significant amount of shared cookies across different domains.