Re: Symmetric alternative to public-key crypto for key-exchange?

arachnidster@gmail.com writes:
> 1) Bob encrypts the key to be exchanged with his key and sends it to
> Alice
> 2) Alice encrypts this message with her key and sends it back to Bob
> 3) Bob decrypts this message with his key, and sends it back to Alice
> 4) Alice decrypts the message, revealing the key

This is called Shamir's 3-pass protocol, independently discovered by
Okamoto.

> This requires a couple of properties from the symmetric cipher used: It
> must be true that D(b, D(a, E(b, E(a, message)))) = message, and it
> must not be possible to recover the key given the intermediate messages
> transmitted between Alice and Bob (for example, with a standard stream
> cipher, XORing subsequent messages would reveal both Alice and Bob's
> keystreams, and hence the original message).
>
> My question is this: is anyone familiar with a cipher that meets these
> criteria?

The obvious one is the Hellman-Pohlig cipher, which does a modular
exponentiation operation similar to various public-key ciphers. That
is, the protocol is about as computationally expensive as a comparable
public-key protocol.
Received on Thu Sep 29 21:51:39 2005