eap Mailing List: RE: Re: Issue 348: Definition of Lower Layer

[Narayanan, Vidya (vidyan]

The definition for lower layer is present in section 2.2 in RFC3748 - I
guess that makes it applicable only between the peer and authenticator.
However, it is a bit confusing to me - I have assumed 3.1 as being
applicable to the link between authenticator and backend server as well.
Since RADIUS is a lock-step protocol, re-ordering is not applicable. If
anything, the requirement on the EAP lower layer here is to be able to
present the packets in order to the EAP layer (the lower layer may
achieve that via re-ordering or by lock-step mechanisms). 

But, given what is present in 3748, I guess this definition would work. 

Vidya


> -----Original Message-----
> From: Bernard Aboba [mailto:bernard_aboba [at] hotmail.com] 
> Sent: Sunday, April 09, 2006 5:00 PM
> To: Narayanan, Vidya; eap [at] frascone.com
> Subject: RE: [eap] Re: Issue 348: Definition of Lower Layer
> 
> I think IKEv2 would qualify, because it carries EAP between 
> the peer and authenticator.  RADIUS (or Diameter) would not, 
> because it carries EAP between the authenticator and backend server.
> 
> Based on the definition of lower layer in RFC 3748, it would 
> seem that Section 3.1 applies only between the peer and 
> authenticator, although some of the same requirements (no 
> re-ordering) would also be relevant between the authenticator 
> and backend server.
> 
> >So, by this definition, RADIUS or IKEv2, for e.g., would not be 
> >classified as a lower layer? I see that the pass through 
> mode figure in
> >RFC3748 alludes to this, but I thought for having the "lower 
> layer" vs.
> >the "AAA" text in the stack was because AAA is more or less 
> standard on 
> >that link, while the "lower layer" between the peer and 
> authenticator 
> >could differ.
> >
> >Section 3.1 of RFC3748 applies to all layers that carry EAP, 
> not just 
> >between the peer and authenticator - isn't that right?
> >
> >Vidya
> 
> 
>