search

freedomofpress / securedrop-protocol

Research and proof of concept to develop the next SecureDrop with end to end encryption.
GNU Affero General Public License v3.0
47 stars 1 forks source link

Compare to and use Oblivious Message Retrieval terminology #35

Open lsd-cat opened 5 months ago

lsd-cat commented 5 months ago

We came to the conclusion that our attempt at not having accounts and using 'challenges' (also named group DH shares and in even other different way across the repo and the issues) is conceptually close to the concept of Oblivious Message Retrieval, recently researched at [1], [2], [3] and more. While we do not directly compare to that, in terms of scale, part of their model is similar and we could use a similar terminology such as 'clues' instead of 'challenges'. The reason why we are not directly using that and we are instead developing our own version, is mainly our use of classic and well tested crypto primitives available in established libraries rather than resorting to fully homomorphic encryption. While I do believe this research area is very promising and a necessary step to hide/reduce even more metadata, it would not be production ready for our timeline. Still a comparison and shared terminology might help a clearer understanding for both cryptographers and general readers.

[1] - https://eprint.iacr.org/2021/1256.pdf [2] - https://link.springer.com/chapter/10.1007/978-3-031-15802-5_26 [3] - https://www.computer.org/csdl/proceedings-article/sp/2024/313000a115/1Ub23ocBmKI

eaon commented 5 months ago

Oh, interesting, looks like my intuition with regards to the Oblivious Transfer terminology was reasonable after all. And despite not knowing about this research, in my work on Reach I've been calling "clues" hints, with a similar justification I suppose.