On the Effectiveness of Intersection Attacks in Anonymous Microblogging

Intersection attacks, which are popular traffic analysis attacks, have been extensively studied in anonymous point-to-point communication scenarios. These attacks are also known to be challenging threats to anonymous group communication, e.g., microblogging. However, it remains unclear how powerful these attacks can be, especially when considering realistic user communication behavior. In this paper, we study the effectiveness of intersection attacks on anonymous microblogging systems utilizing Twitter and Reddit datasets. Our findings show that the attacks are effective regardless of whether users post their messages under pseudonyms or publish them to topics without attaching identifiers. Additionally, we observed that attacks are feasible under certain settings despite increasing userbase size, communication rounds' length, cover traffic, or traffic delays