[WIP] pake crate

[tarcieri]

Traits which describe the functionality of Password-Authenticated Key Exchange algorithms, which can be used to establish a shared secret over an insecure channel using only the client's knowledge of a password and the server's knowledge of a password hash for authentication.

[rozbb]

Why is "proof" used here? Usually we just refer to message 1 and message 2 for a 2-message PAKE. Here's an example of 3 popular PAKEs

I wrote them as one round, but it doesn't matter implementation-wise, since usually one side is the designated initiator anyway.

[tarcieri]

@rozbb happy to make the language more generic. The terminology there was coming from SRP which is what I was working on at the time (mostly I wrote this up trying to think of how to refactor the existing srp crate API into something more manageable)

[rozbb]

Ok nice. Also, I assume you only want to include 2-message PAKEs for now? Those are the only standardized one atm.

I intend on implementing CPace in Rust soon, so I will be able to have informed feedback on this.

[tarcieri]

Yes, that sounds good. FWIW we have AuCPace

[makavity]

Hey, @tarcieri. Since we've released SWU function, my next plan is to do bpace protocol. Should I wait for the traits, or I can PR it right now?

[tarcieri]

@makavity I'd say work on this is stalled for now, so don't let it be a blocker.

I intend to pick this up again after finishing stable releases of everything else.