Privacy-Preserving Contact Discovery with Applications to End-to-End Encrypted Messaging and Mobile-First Cryptocurrencies
Please download main.pdf to read the full report.
In this report we consider means to perform privacy-preserving contact discovery on mobile devices. Contact discovery is a crucial initial step for any social application. Current methods reveal private information or require users to reason about cryptography. We introduce an approach based on non-interactive identity-based key exchange protocols (NI-IBKE). Our scheme provides users with cryptographic material which is entirely managed by a client-side application. Users can input their contacts’ phone numbers (or an equivalent human-readable identifier) to compute shared secret keys on-device. Our system relies on a t-out-of-n trust assumption with respect to a decentralised service. We show that the desired privacy property holds in the random oracle model under the decisional bilinear Diffie-Hellman assumption. We provide estimates of the system’s performance in the wild. These show that our scheme is applicable but will require optimisations or a long bootstrapping period if used for applications with billions of users. Finally we describe a simple proof-of-concept implementation as well as applications of the scheme for end-to-end messaging or mobile-first cryptocurrencies.
See proof-of-concept implemented in Go using the kyber library.