On Nov 25, 2008, at 3:32 PM, Sumanth Channabasappa wrote:

> Scott,
>
> I had additional offline discussions with Ekr and Cullen last week.  
> Ekr
> indicated that the proposed I-D does not align with the SIP security
> model (UE <=> next-hop authentication; no need for UE <=>
> <authentication beyond next hop>).
>

Huh?

We have long held that there may be multiple proxies (possibly in  
different domains) challenging a given request. This has been a  
hallmark of scenarios such as the "hotel proxy" that doesn't do  
authentication or act as a an identity server but that does do local  
firewall control.

So somebody is confused about something. Who knows, might be me,  
occasionally I wake up in a different time-space continuum, but AFAIK,  
we've always intended to support authentication more than one hop  
away. In fact, I recall design discussions with Pingtel and 3Com  
people about that as far back as 1999. Robert Sparks had a lot to say  
about it, IIRC.

[S] My understanding -- based on the offline discussions -- is that the
proxy can still challenge the UA. However, the UA trusts its connection
with the next-hop and does not need to authenticate any elements beyond
the next-hop. In other words, mutual authentication is not necessary (I
used <=> to indicate mutual authentication). [[ Today, we allow for the
Authentication-Info header that allows the UA to authenticate the
registrar (even after it has established TLS with the next-hop proxy),
which also breaks this model. ]]

- S




_______________________________________________
Sip mailing list  https://www.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use [EMAIL PROTECTED] for questions on current sip
Use [EMAIL PROTECTED] for new developments on the application of sip

Reply via email to