CatNap: Leveraging Generic MPC for Actively Secure Privacy-enhancing Proximity Testing with a Napping Party

Paper i proceeding, 2022

Proximity testing is at the core of several Location-Based Services (LBS). Despite a series of reported and confirmed abuses, modern LBSs still demand their clients to disclose their locations in plain in order to preform location proximity testing. This works aims at enhancing proximity testing with privacy. We design CatNap a novel protocol that (1) implements precise Euclidean distance matching; (2) allows matching even if the clients are not online at the same time (the "napping party" feature); (3) is secure against active adversaries (malicious actors that corrupt up to one party); (4) makes black-box use of generic Multi-Party Computation techniques (any future improvement of the underlying building blocks will also boost CatNap); and (5) is efficient: servers run with about 0.03 seconds of CPU time and 5.6MB of communication, while clients perform only a small number of Boolean operations and need just 51 bytes of communication.