Re: channel binding
From: Yoshihiro Ohba (yohbatari.toshiba.com)
Date: Thu, 1 Sep 2005 00:17:13 -0400 (EDT)
On Wed, Aug 31, 2005 at 09:12:28PM -0700, Salowey, Joe wrote:
> > 
> > The draft-ohba-eap-aaakey-binding-01 draft assumes that a 
> > blob carries static information only.  Thus it is possible 
> > for the AAA server to have the information to validate the 
> > blob itself or a hash of the blob.  If this is the case, 
> > carry a hash of parameters in RADIUS is sufficient.
> 
> [Joe] It is sufficient in the case when there is no variability in what
> the authenticator will send.  In this case why send it at all?  

Good point.  If the authenticator sends it, the AAA server does not
need to have the information pre-configured (i.e., ease of
management).  Of course this works only when the authenticator is
fully trusted by the server.

> If there
> is variability in what the authenticator will send then it may not be
> straight forward to validate what the authenticator sends if it is just
> a hash value.  

Yes, this would be true.  On the other hand,
draft-ohba-eap-aaakey-binding-01 does not assume the variability.

Yoshihiro Ohba

Results generated by Tiger Technologies using MHonArc.