Re: Is symmetric key distribution equivalent to symmetric key generation?

From: Alan (a__l__a__n_at_hotmail.com)
Date: 08/23/05


Date: Tue, 23 Aug 2005 11:06:03 -0400

chuckles wrote:
> Would you place the diffie hellman algorithm in the same class as
> distribution of symmetric keys using asymmetric cryptography?

The biggest difference is that a pure Diffie Hellman key excange is not
authenticated.

Diffie Hellman enables two people who never met to arrive at a shared secret
key which neither of them controls and neither could know in advance. But
DH is vulnerable to man-in-the-middle attacks. (A thinks he is performing
DH with B, and B thinks he is performing it with A. In reality, there is a
third party C in the middle performing the key exchanges with each of A and
B. So C decrypts all messages, reads and perhaps alters them, and
re-encrypts to send on to the intended party). To avoid this problem you
would need some sort of authentication in the protocol (google for
"authenticated diffie hellman" for some papers on the subject).

Public (asymmetryc) key encryption provides a means to exchange a secret key
in an authenticated way, PROVIDED that each party has a trusted copy of the
other party's public key. You would need to supplement it with something
else if you need a shared secret key which is not controlled by either party
and not known to either party in advance.



Relevant Pages