Ad-Hoc Secure Two-Party Computation on Mobile Devices using Hardware Tokens

Secure two-party computation allows two mutually distrusting parties to jointly compute an arbitrary function on their private inputs without revealing anything but the result. An interesting target for deploying secure computation protocols are mobile devices as they contain a lot of sensitive user data. However, their resource restriction makes the deployment of secure computation protocols a challenging task. In this work, we optimize and implement the secure computation protocol by Goldreich-MicaliWigderson (GMW) on mobile phones. To increase performance, we extend the protocol by a trusted hardware token (i.e., a smartcard). The trusted hardware token allows to pre-compute most of the workload in an initialization phase, which is executed locally on one device and can be pre-computed independently of the later communication partner. We develop and analyze a proofof-concept implementation of generic secure two-party computation on Android smart phones making use of a microSD smartcard. Our use cases include private set intersection for finding shared contacts and private scheduling of a meeting with location preferences. For private set intersection, our token-aided implementation on mobile phones is up to two orders of magnitude faster than previous generic secure two-party computation protocols on mobile phones and even as fast as previous work on desktop computers.