You are rapidly converging on the real thing. Secure access is not
necessary for a client, and should never be made a required part of the
spec, using the the FNP sessions does not complicate things for normal
client developers.

On Tue, Jun 26, 2001 at 07:22:06PM -0700, Ian Clarke wrote:
> On Tue, Jun 26, 2001 at 10:22:05PM -0400, Benjamin Coates wrote:
> > As long as we have a secure way of negotiating the FCP port and don't rely 
> > on 
> > an assumed default port.  (it'd be too easy for another local user to 
> > listen 
> > on the default port and spoof a FCP password request)
> 
> Hmmm, good point.  This suggests that a slightly more robust
> challenge-response approach would be preferable, at the risk of
> increasing the difficulty of client authorship.  With the addition of a
> secure hashing algorithm to the client (SHA1 anyone?) this can be
> achieved.  The protocol is:
> 
> 1) Client connects to FCP port on node
> 2) node sends random string to client
> 3) client appends random string to plaintext password and hashes result
> 4) node does same
> 5) client sends hash to node which compares its hash to the one created
>    by the client, and if they are the same the client is authenticated
> 
> Perhaps there is an easier hashing algorithm (from an implementation
> standpoint) that we could use in place of SHA1 which would also be
> secure.
> 
> Ian.



-- 
'DeCSS would be fine. Where is it?'
'Here,' Montag touched his head.
'Ah,' Granger smiled and nodded.

Oskar Sandberg
oskar at freenetproject.org

_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to