Re: Diffie-Hellman for Matchmaking?

[lmccarth@cs.umass.edu]

Dimitris Tsapakidis writes:
> Person A is interested to match person B, so he computes
> g^(AB)mod n. B is interested in X, where X may or may not
> be A, and calculates g^(BX)mod n. Now, they compare these
> two "common keys" either using some Zero Knowledge scheme
> that ensures fairness (at no point one party has significantly
> more information than the other) or through a Trusted Third Party.
> If they are the same, then this means X=A, so A and B
> have a match (e.g. a date). The common keys must remain
> secret (hence the ZK above): if g^(BX)mod n "escaped"
> to the public, then the real X would find out that
> B is interested in him.

Could you give us some background on the problem ?  I'm not clear on what
the protocol is trying to achieve in practical terms.

-Lewis