Hi Sam:

Let me answer this one as well.

El 10/11/2011, a las 15:47, Sam Hartman escribió:

>>>>>> "Alejandro" == Alejandro Perez Mendez <[email protected]> writes:
> 
>>> 
>    Alejandro> on the first place, to make the KDC Moonshot-enabled you
>    Alejandro> need it to support GSS preauth.
>>> 
>>> Or define an EAP preauthentication fast factor.
> 
>    Alejandro> I guess you meant GSS-EAP preauthentication fast factor,
>    Alejandro> right?
> 
> No.
> Any preauth mechanism is a FAST factor.
> So if there is a GSS preauth there is a GSS and thus GSS-EAP FAST
> factor.
> 
> I mean go back to a model where it's EAP between the RP and KDC not GSS.

So indeed we would be defining a new type of backend transport protocol for EAP 
based on Kerberos

> 
> I've been thinking more about this.  It's obvious to me that we don't
> want to use the same GSS-EAP context between the client and RP that we
> do between the RP and KDC, even if we use the same EAP conversation.  My
> rationale is a bit complicated. One of the big reasons is that the
> question of which enctype to use, what GSS-EAP features to use etc
> should depend on the RP capabilities not on the KDC capabilities.  Also,
> (and perhaps this is what Nico's getting at) it simplifies GSS channel
> binding between the client and RP.
> 
> This doesn't mean we can't use GSS-PREAUTH; it just means using
> GSS-PREAUTH is kind of complicated.
> So, like Josh, I'm beginning to ask what GSS buys us in
> preauthentication?

Let me clarify one aspect about GSS-PREAUTH in Kerberos 
(http://tools.ietf.org/html/draft-perez-krb-wg-gss-preauth-00). The idea we had 
when designing GSS preauth in Kerberos is to provide a flexible 
pre-authentication mechanism for Kerberos based on the multiple mechanism that 
GSS-API could offer (in particular, the GSS-EAP mechanism can be used as we 
described in http://tools.ietf.org/html/draft-perez-abfab-eap-gss-preauth-00). 
In fact, this flexibility was one the advantages we mentioned with respect to 
the use of EAP packets directly as pre-authentication data in Kerberos, which, 
by the way, we also described some time ago 
(http://www.ietf.org/mail-archive/web/abfab/current/msg00040.html). 

Obviously, the goal of our work is to define a flexible pre-authentication 
mechanism for Kerberos, but not a transport for EAP in the backend side of the 
acceptor/authenticator. However, I understand that if the same concept might be 
used it would really nice but, as I said, goals are completely different under 
my point of view.

Best regards.


> _______________________________________________
> abfab mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/abfab

-------------------------------------------------------
Rafael Marin Lopez, PhD
Dept. Information and Communications Engineering (DIIC)
Faculty of Computer Science-University of Murcia
30100 Murcia - Spain
Telf: +34868888501 Fax: +34868884151 e-mail: [email protected]
-------------------------------------------------------




_______________________________________________
abfab mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/abfab

Reply via email to