Fighting Spam in Dating Apps

Abstract

Online dating allows for interactions between users with a high degree of connectivity. This digital form of interaction attracts spammers and scammers, who try to trick users to visit low-class competitors’ websites or steal the users’ money. Fortunately, each attacker leaves a footprint of its actions in the network. It is the task of an Anti-Spam system to detect these and punish the culprit. In this paper, we demonstrate how LOVOO fights such illegal activities using a system of modular, scalable, and fault-tolerant Anti-Spam components. We describe our stream-processing architecture and how it ensures flexibility and facilitates resource-efficient processing of the millions of events produced by hundreds of thousands of users.