Re: FW: I-D Action:draft-kaplan-sip-session-id-00.txt

On Mon, 2008-12-08 at 13:27 -0500, Hadriel Kaplan wrote:
> > I'm still troubled by what I perceive as an inversion of the names: I
> > think of a 'session' as something that happens within a 'call'.  I
> > believe that you're looking for something that's a parent of 'call-id's,
> > not a child.  Might I suggest 'Call-Set-Id'?
> 
> Yeah, header names are funny things.  We want them to mean X, but
> other people interpret them to mean Y, and in reality they end up
> meaning Z. :)
> 
> I used the term "Session-ID" because it was what I thought common
> operators of SIP deployments would expect it to be, because I think
> most people think of a Session as actually a superset of a Call-ID -
> the Call-ID could change 15 times along the path, but it's the same
> session. (in many ways a Call-ID is not a "Call" identifier at all!)
> But I grant that it confuses us with media "sessions".
> 
> How about: "Correlation-ID", or even "Opaque-Call-ID"?

It's treacherous -- the original IETF definition was for "session",
which meant a particular stream of media bytes.  SIP is "Session
Initiation Protocol", a wrapper around sessions.  But with the use of
re-INVITE, one can change the enclosed session in a myriad of ways.

If we're talking about a "sequence" of SIP dialogs which are connected
end-to-end by B2BUAs, then they all enclose the *same* session.  (Which
suggests the common identifier should be put into the SDP!)  But if we
use the identifier to group dialogs in any other way -- and there have
been suggestions to do so -- then we can't call the group of dialogs a
"session" without creating confusion.

Dale


_______________________________________________
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