> -----Original Message-----
> From: Tschofenig, Hannes (NSN - FI/Espoo)
> [mailto:[email protected]]
> Sent: Thursday, March 15, 2012 12:11 AM
> To: ext Jim Schaad; [email protected]; Hannes Tschofenig
> Cc: [email protected]
> Subject: RE: [abfab] draft-jones-diameter-abfab-00 question
> 
> Hi Jim,
> 
> > I will state upfront that I know less about diameter than I do about
> > Radius so the questions I have are to be taken with that grain of
> > salt.
> >
> > 1.  Is it possible to include a more general SAML query than just an
> > authorization request in a DER message?  Specifically, I would like to
> > be able to query for a set of attributes about the entity that was
> > authorized as oppose to get the fact they are authorized.
> 
> Yes, this is possible.
> 
> Some problems we are dealing with in RADIUS, such as the fragmentation
> issue, does not exist in Diameter due to the TCP transport.

Ok - I am slightly confused.  When I look at the answer ABNF - it has both a
SAML-AuthnResponse and a SAML-Assertion.  However the request ABNF - it only
has SAML-AuthnRequest but not the SAML-Assertion.   I read this as saying
that one should/could not include the SAML-Assertion at this point.
> 
> >
> > 2.  Does Diameter give any way of sending the keys around that are to
> > be used for doing the xml encryption operation?  I understand that
> > diameter is more point-to-point than RADIUS but I do not know that to
> > be a fact.
> > Does
> > this mean that there is more likely to have end-to-end signing and
> > encryption capabilities present?
> 
> Diameter also exchanges keys in the same manner as RADIUS does. This is,
> for example, used for the Diameter EAP application. These keys are,
> however, secured only in a hop-by-hop fashion (and not end-to-end; the
> ends are Diameter client and Diameter server).
> 
> Currently, there is no standard for e2e AVP encryption in Diameter
natively
> although there are discussions ongoing to do some work in the DIME working
> group. Two relevant contributions for that matter are:
> 
> http://datatracker.ietf.org/doc/draft-korhonen-dime-e2e-security/
> http://datatracker.ietf.org/doc/draft-zorn-dime-n2n-sec-lite/
> 
> (Note that the scope of the two drafts is a bit different.)
> 

But back to my original question - where would the xml encryption process
get it's keys?


Jim

> >
> > 3.  Is there a concept of proxies that sit on boundaries that could
> > modify the SAML constructs to deal with mapping of attributes?
> 
> There are proxies in Diameter but what they specifically do with certain
AVPs
> depends very much on the application. The document I have submitted
> describes such an application but does not describe such operation.
> 
> Ciao
> Hannes
> 
> >
> > Jim
> >
> >
> > _______________________________________________
> > abfab mailing list
> > [email protected]
> > https://www.ietf.org/mailman/listinfo/abfab

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

Reply via email to