Re: [sipcore] draft-roach-sipcore-rfc3265bis-00

["Dale Worley" <[e-mail]>]

In regard to the fact that a proxy (that wishes to remain in the
signaling path) must Record-Route all SUBSCRIBEs without to-tags, and
all NOTIFYs:

On Tue, 2009-08-18 at 15:24 +0200, DRAGE, Keith (Keith) wrote:
> Christer wrote
> 
> > Again, a 3261 compliant proxy may Record-Route the SUBSCRIBE, 
> > but that doesn't mean it will Record-Route the NOTIFY. It may 
> > simply be configured to Record-Route all dialog initiation requests.
> > 
> 
> It can only do this if it understands that SUBSCRIBE is a dialog
> initiation request, and it can only do this if it implements either
> RFC 3265 or 3265bis. 
> 
> As I understood it, the behaviour described by Adam is already in RFC
> 3265, and therefore it is a non-conformant implementation of RFC 3265
> if it does something difference.

Some years ago, I was in a long discussion about how SUBSCRIBE/NOTIFY
need to be record-routed, especially in the context of
non-dialog-stateful proxies.  Unfortunately, nobody in the conversation
recalled section 3.1.5 of RFC 3265, but we came to the conclusion that
in order for subscriptions to work *in practice*, a non-dialog-stateful
proxy must record-route SUBSCRIBEs without to-tags and all NOTIFYs, or
record-route none of them.  So practical demands require that any proxy
in an environment that uses SUB/NOT would satisfy 3.1.5 (in almost all
cases), that is *be* RFC 3265-compliant, even if it didn't *intend* to
be RFC 3265-compliant.

In that context, 3265bis doesn't demand any behavior from proxies that
they don't (in practice) have to have already.  So I think we can safely
depend on them having the proper behavior.

Dale