Re: [SECMECH] Framework Bindings Vs. Mechanism Bridges

On Fri, Aug 26, 2005 at 08:06:48AM -0700, Salowey, Joe wrote:
> > From: Bernard Aboba [mailto:aboba@internaut.com] 
> > Sent: Thursday, August 25, 2005 11:40 PM
> > 
> > It's different because the NAS needs to support Kerberos even 
> > if it is operating in "Pass-through" mode.  
> 
> [Joe] Then its not operating in "Pass-through" mode, it is terminating
> EAP (acting as an EAP server).

I agree.

As for fast reconnect/handoff, which I'd mentioned earlier, the re-use
of TGTs and service tickets (within their lifetimes), can save round
trips in the IAKERB method, but doesn't really amount to fast
reconnect/handoff in pass-through mode in that the method must be used
everytime and so the EAP server is involved always.

What is the exact definition of "fast reconnect/handoff" anyways?
RFC3748's description of "fast reconnect" does not say very much.

> > In contrast, a 
> > NAS operating in pass-through mode for EAP-TLS doesn't need 
> > to validate the client certificate. 
> > 
> > > It could also be possible to still involve a AAA and have 
> > it terminate 
> > > the method and talk to the KDC. If you are trying to implement a 
> > > method that is evaluated by both the NAS and the AAA in the same 
> > > transaction you are really doing something other than EAP.
> > 
> > In all the EAP Kerberos proposals I've seen the method is 
> > terminated on either the NAS or AAA server, but not both.  
> > But in any scenario, the NAS still needs to support Kerberos, 
> > in order to validate the "network access" service ticket. 
> 
> [Joe] If this is done in the same method execution instance it seems
> problematic to me, because you have in effect two EAP-servers (one in
> the AAA and one in the NAS) processing the EAP messages.  This is not
> EAP. 

I agree.  If the NAS speaks Kerberos V and, more importantly, if there's
a cross-realm path from the peer's principal's realm to the NAS's realm,
then NAS is (should be) the EAP server also; if the NAS does not speak
Kerberos or there is not such x-realm path then, as far as the NAS is
concerned, the EAP Kerberos V method is like any other AAA method.

Nico
-- 

_______________________________________________
SECMECH mailing list
SECMECH@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/secmech