Thank You For Being A Friend: An Attacker View on Online-Social-Network-Based Sybil Defenses

Online Social Networks (OSNs) have become a rewarding target for attackers. One particularly popular attack is the Sybil attack, in which the adversary creates many fake accounts called Sybils in order to, for instance, distribute spam or manipulate voting results. A first generation of defense systems tried to detect these Sybils by analyzing changes in the structure of the OSN graph---unfortunately with limited success. Based on these efforts a second generation of solutions enriches the graph-structural approaches with higher-level user features in order to detect Sybil nodes more efficiently. In this work we provide an in-depth analysis of these defenses. We describe their common design and working principles, analyze their vulnerabilities, and design simple yet effective attack strategies that an adversary could launch to circumvent these systems. In our evaluation we reveal that an miscreant can exploit the credulity of OSN users and follow a targeted attack strategy to successfully avoid detection by all existing approaches.