> > Cullen said transitive trust of the entire SIP network for 
> > Identity was fine
> 
> Uh, I don't think I exactly said that but if I did, I like to  
> contradict myself now :-) Actually if I thought this, it seems that  
> the From header would be fine and we would have no need for 4474 or  
> 3325.

RFC4474 and RFC3325 are useful and valuable and necessary.  I have
not said otherwise.  Thing is, their applicability on the SIP networks
being designed and deployed is limited, much as SIP's early design
had UACs sending SIP directly to UASs and slowly proxies became
necessary due to NATs, firewalls, network-provided QoS (via SDP
snooping), and so on.

I'm not sure why you removed ", and RFC4474 was sufficient even with 
SBCs in the network" from my sentence, as it was relevant to the
meaning of my statement.  As I explained in Chicago in the last timeslot
in SIP, RFC4474 doesn't work if the Invite traverses SBCs that are
not owned by the originating domain's authentication agent.  SIP 
networks deployed and operating today have SBCs that aren't operated 
by the originating domain's authentication agent.  Thus, I don't 
see how we get identity in a network with SBCs, except via 
transitive trust:  I trust my provider to have validated the
signature they received, who trusts the upstream provider to have
validated the signature they received, and so on.  This has 
security properties identical to using P-A-I between those 
domains.

draft-ietf-sip-answermode is only one example of the significant 
need for identity in SIP.  Without identity that works across 
SIP networks being built today, we can't provide the foundation 
for its necessary authorization mechanism.

-d


_______________________________________________
Sip mailing list  https://www1.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