[859] in cryptography@c2.net mail archive

home help back first fref pref prev next nref lref last post

Re: forward secrecy and email protocols

daemon@ATHENA.MIT.EDU (Peter Gutmann)
Fri May 16 17:26:06 1997

From: pgut001@cs.auckland.ac.nz (Peter Gutmann)
To: hal@rain.org
Cc: cryptography@c2.net
Reply-To: pgut001@cs.auckland.ac.nz
X-Charge-To: pgut001
Date: Sat, 17 May 1997 09:05:33 (NZST)

>One of the ideas we talked about on cypherpunks a while back was
>to keep a session key around for each person you communicate with.
>For each message, you use the session key, and then you run it through a
>one-way function to derive the session key to use for the next message.
>There would need to be an initial public key communication to set up
>the session keys, but if the eavesdroppers didn't have access to that
>(because it happened a long time ago) then you get forward secrecy.
 
You don't really get forward secrecy because if someone compromises session 
key n then they also have n+1, n+2, ...  A while back I thought of using 
something like SKEME to exchange a master secret and derive new session keys 
from this (that is, from the master secret, not from another session key), 
which still isn't perfect because a compromise of the master secret (rather 
than one session key) compromises the session keys.  A problem with SKEME (and 
Oakley I think) is that the "fast rekey" mechanism is based on the initial key 
shared via PKC-encrypted messages rather than the DH-derived session key, so a 
compromise of the private keys also gives you all the session keys.  What 
would be nice is a cheap (ie less expensive than full DH) method of agreeing 
on a new session key based on an existing shared secret which doesn't weaken 
the new key if the secret is later compromised.  SKEME offers something which 
is a step in this direction by allowing you to do a new EXCH step during idle 
time and authenticating it when you need a new key with an AUTH step, which at 
least moves the DH cost into the background.
 
(Speaking of SKEME and Oakley, I haven't heard much of these recently - have 
there been any developments with them since late last year?  SKEME seems far 
too elegant to sit around not being used...).
 
Peter.
 


home help back first fref pref prev next nref lref last post