Re: [Sip] draft-ietf-sip-gruu-10 comments

[Jonathan Rosenberg <jdrosen at cisco.com>]

My view on this is that, with GRUU, the home proxy in the terminating 
domain is going to get visited ANYWAY, even with just a trapezoid. So 
there is no penalty to record-routing, besides perhaps the overhead in 
the sip message for the extra bytes. Thus it seems simpler to say that 
you should always do it.

It is a fair point that in mixed mode deployments where not all 
endpoints in a domain support gruu, the record-route would not be needed 
and would introduce an extra hop. I think it would be OK to say that if 
the registered contact was assigned a gruu, then the home proxy has to 
record-route. Would that be OK?

-Jonathan R.

Cullen Jennings wrote:

> On Sep 15, 2006, at 1:42 PM, Paul Kyzivat wrote:
>
> > Cullen Jennings wrote:
> >
> > > On Sep 11, 2006, at 12:21 PM, Paul Kyzivat wrote:
> > >
> > > > Cullen Jennings wrote:
> > > >
> > > > > Section 8.2.2
> > > > > I think I am likely confused here. I basically don't see why a  
> > > > > simple classic sip trapezoid deployment would ever need to  Record 
> > > > > Route. The fact that the GRUU refereed to the "home"  domain would 
> > > > > cause it work get back here. As a side note, I find  the terms 
> > > > > "home", "edge" etc very undefined outside the IMS  context. 
> > > > > Anyways, I believe this section might be all right but  I am 
> > > > > failing to understand what needs to be done and why. If I  was a 
> > > > > non IMS proxy implementer, I suspect I would just ignore  this 
> > > > > whole section.
> > > >
> > > >
> > > > *If* the topology is a simple trapezoid then I think you are  correct.
> > > >
> > > > But if there is a proxy before the originating home proxy, or  after 
> > > > the terminating home proxy, then the R-R is needed to  prevent a 
> > > > spiral.
> > > >
> > > > On the originating side, this can be determined by whether there  is 
> > > > a R-R present when the request arrives at the originating home  
> > > > proxy. On the terminating side that doesn't work, because the  
> > > > potential R-R hasn't happened yet, and you can't know in advance  if 
> > > > it will happen.
> > > >
> > > > So the problem is that you don't need it in a trapezoid, but you  
> > > > don't know if you have a trapezoid.
> > > But on the terminating side, it seems like the home proxy know  there 
> > > is another proxy because it needs to insert that into the route
> > > (I don't spend enough time thinking about this so I assume you are  
> > > right, just trying to make sure I understand)
> >
> >
> > Well, the terminating home proxy will know if it is including  another 
> > proxy after it. But it doesn't know if that proxy, or some  other 
> > proxy *it* includes, will R-R.
> >
> > And even if the home proxy isn't inserting a proxy (e.g. no Path  
> > header registered), it doesn't really know if the contact that was  
> > registered is the actual destination, or whether it identifies  
> > another proxy.
>
>
> Ok, but that seems like the advice should be "if you are using the  path 
> stuff (or the broken 3rd party registrations) to include another  proxy, 
> then you must R-R".  Making everyone R-R for no good reason  just seems 
> like a complicated wasted and will slow down systems that  don't need it.

--
Jonathan D. Rosenberg, Ph.D.                   600 Lanidex Plaza
Cisco Fellow                                   Parsippany, NJ 07054-2711
Cisco Systems
jdrosen at cisco.com                              FAX:   (973) 952-5050
http://www.jdrosen.net                         PHONE: (973) 952-5000
http://www.cisco.com

_______________________________________________
Sip mailing list  https://www1.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use sip-implementors at cs.columbia.edu for questions on current sip
Use sipping at ietf.org for new developments on the application of sip