Membership-Private Registration-Based Encryption

Aditya Hegde

Video

Abstract:

Registration-based encryption (RBE) lets senders encrypt to a user identity using publicly maintained registration state, without requiring a trusted key-generation authority. Existing RBE constructions, however, leave open a fundamental privacy gap: while they may hide message contents and sometimes the ciphertext recipient, they do not hide the set of registered users from other clients. This is a serious limitation for private contact-discovery and messaging applications, where users may not wish to reveal the fact that they have registered. In this work, we study membership privacy, which requires that the RBE scheme hide which identities are registered, even against adversaries that adaptively register arbitrary users and obtain update information for any user. We demonstrate that existing RBE schemes fail to achieve membership privacy and present membership-private RBE constructions for large-universe identity settings. We show that our schemes are practical for real-world deployment: with one million registered users, our implementation achieves ciphertexts of only a few tens of kilobytes and amortized registration times on the order of tens of milliseconds. We further provide estimates for up to one billion users, showing that ciphertexts remain within a few tens of kilobytes and the estimated amortized registration time is on the order of a few seconds. This is joint work with Alexis Cao and Matthew Green.

Bio:

Aditya Hegde is a PhD student in Computer Science at Johns Hopkins University, advised by Abhishek Jain and Matthew Green. His research spans theoretical and applied cryptography. Much of his work focuses on secure computation and, more recently, on designing primitives with low communication and minimal interaction.

Time and Place

Thursday, November 12, 4:00pm
CoDa E201 & Zoom