Mahesh Anjanappa wrote:
I'm not fully understanding what you are proposing (see questions above). I expect that it may start to fall apart in the details of particular use cases. Will you explain to me how your definition relates to the use case with A/B/C/D that I gave?

OK,let me see.
A, C and D get into a Session/Contract to communicate using Video and Audio.

I guess you misunderstand what I was describing.

The SDP negotiated between A and B contains both audio and video.

The SDP negotiated between B and C contains only audio. (B didn't offer the video to C.)

The SDP negotiated between B and D contains only video. (B didn't offer the video to D.)

The purpose of B is to consolidate an audio-only device and a video-only device into what seems, to A, like one audio/video device.

They don't all have the same contract. So by your definition I don't think they can all have one session.

In some conceptual sense you may think of this as one session. But its hard to find any tangible item that they all have that refers to a single session. In the end, from the point of view of B it may indeed be one session. But A, C, and D can't perceive that.

The contract allows for either of the Media to be used hence C uses audio
and D uses video. B facilitates the Contract by holding the three dialogs.
The 3 dialogs are tied together by a Conference Id (i suppose)
thus COnference Id representing the Session Identifier.B too is considered a party in the
session though its role is different from that of A,C and D.

If this were done with a conference mixer, and hence there was a conference id, then I guess you could say that the conference id was an identifier for the "session".

But if B was purpose-built for this particular functionality, and there is no conference id, then what?

And in a simple call between A and C there would be no conference id, so something else must serve as the "session" identifier. But in that case, it would seem that these aren't the *same* notion of session.

You asked why I didn't think it was important to define session. I will admit that it seems odd to have a Session Initiation Protocol without a clear idea of what this "session" is that it is initiating. So it indeed would be *nice* to have a definition that covers all the cases.

But not having one doesn't seem to be preventing anyone from doing anything, does it?

From a practical perspective I think the thing that is being initiated by SIP is the contract established by a matching offer/answer. That is fairly straightforward when the UAC and UAS are the offerer and answerer, or visa versa. It gets messy when other UAs are involved in the construction of the offer and the answer. And in those cases it gets really hard to figure out what the session is, because not all the parties will agree.

        Paul

Paul

regards
Mahesh

----- Original Message ----- From: "Paul Kyzivat" <[EMAIL PROTECTED]>
To: "gaurav katiyar" <[EMAIL PROTECTED]>
Cc: <[email protected]>; <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Tuesday, April 29, 2008 10:32 PM
Subject: Re: [Sip] [SIPForum-discussion] What's the difference between session and dialog in SIP?




gaurav katiyar wrote:
Sip "Session Initiation Protocol". It talks about media session, i mean
how to initiate, modify and terminate a session and Dialog provides a
sort of context (like caller, callee, location, routing etc.) to modify
this session. So dialog is more related to route the sip message to
right node and body (SDP) handles media session if present. Dialogs may
have media sessions Or may not.

Take an example of B2BUA. It is a type of statefull proxy having two
dialog and one media session if doing media routing.

Maybe. IMO these concepts aren't sufficiently defined.

One view is that a session is whatever SDP describes. Thus its one
session regardless of how many media streams it contains. And in the
case of an SDP containing multicast media addresses, and possibly
advertised via SAP, there could be many participants in the same session.

But then in sip it takes a pair of SDPs to establish a call. Is that one
session, or two? And in your case above with the B2BUA, is it one
session, or two, or four?

Another view is that a session in the context of sip is whatever is
established as a result of an INVITE. Then I guess the pair of SDPs
together describe one session, regardless of how many media streams. But
in that case the B2BUA example would be two sessions.

Then consider a more complex case:

                    audio
                 /-------- C
                /
  A ========== B
    audio       \
    video        \-------- D
                    video

Here B is a B2BUA and media relay as in your example, but one stream is
relayed to C and the other to D.

In this case, how many sessions are there? A, C, and D each have one,
but are any of those the same? Or are there three altogether?

When I see "session" used, I just assume it means something vague, and
read on to see what is actually meant.

Paul

*/Donald Lee <[EMAIL PROTECTED]>/* wrote:

    old good questions in sip, also add another "transaction".


    On Fri, Apr 25, 2008 at 8:18 AM, 孙宗� <[EMAIL PROTECTED]
    <mailto:[EMAIL PROTECTED]>> wrote:

        Hi, Vijay

        Thanks very much for your timely answers.

From this specification, session is a combination of signaling
        plane
        and media plane messages and processes that enable two or more
participants to communicate. I think session is a larger scope than
        dialog. In another word, one session can contain more than one
        dialog.
        Is what I understand right?

        Zongjun


        2008/4/24, Vijay K. Gurbani <[EMAIL PROTECTED]
        <mailto:[EMAIL PROTECTED]>>:
         > There is some work progressing in the BMWG WG to create
        performance
> metrics around SIP. One of the first tasks in such an endeavor > is to define the notion of dialogs and sessions. Please see
         >  Section 3.1.1 of
         >

https://svn.resiprocate.org/rep/ietf-drafts/gurbani/bmwg-sip-bench-term/draft-ietf-bmwg-sip-term-00.txt
         >  of the above draft for some insight.
         >
         >  Ëï×Ú¾ý wrote:
         >
         > > Hi, Lincoln
         > >
         > > thanks for your instructions.
         > >
         > > I know that dialog can exist without session since
        SUBSCRIBE/REFER can
         > > create dialog and without any media between communication
        peers.
         > >
         > > What I want to know is the relationship of dialog and
        session when
         > > they both exist in one communication activity. Take a
        example, when
         > > there are 2 person participating talks with voice, we say
        there is a
         > > dialog and a session.
         > >
         > > But when one caller invites another callee and gets five
        200 final
         > > responses from 5 UA every of which has its own session
        description. We
> > can say that there are 5 dialogs between caller and the other 5
         > > callees, right? And then what is the exact number of
        session in this
> > scenario? One session or five session? That is what I want
        to know.
         > >
> > Dialog is determined by dialogID (call-id, from/to tag) and
        session id
         > > is determined by session id given in the SDP message.  My
        answer is
         > > there are 5 dialog and one session now, right?
         > >
         >
         >  - vijay
         >  --
         >  Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
         >  2701 Lucent Lane, Rm. 9F-546, Lisle, Illinois 60532 (USA)
         >  Email: [EMAIL PROTECTED]
        <http://alcatel-lucent.com/>,bell-labs.com
        <http://bell-labs.com/>,acm.org <http://acm.org/>}
         >  WWW:   http://www.alcatel-lucent.com/bell-labs
         >
        _______________________________________________
        Sip mailing list  https://www.ietf.org/mailman/listinfo/sip
        This list is for NEW development of the core SIP Protocol
        Use [EMAIL PROTECTED]
        <mailto:[EMAIL PROTECTED]> for questions on
        current sip
        Use [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> for new
        developments on the application of sip




    --     BR
    Donald _______________________________________________
    This is the SIP Forum discussion mailing list
    TO UNSUBSCRIBE, or edit your delivery options, please visit
    http://sipforum.org/mailman/listinfo/discussion
    Post to the list at [EMAIL PROTECTED]




Gaurav Katiyar
Induslogic india pvt. ltd
B-34/1 sector 59 NOIDA
Phone:9818381368


------------------------------------------------------------------------
Be a better friend, newshound, and know-it-all with Yahoo! Mobile. Try
it now.
<http://us.rd.yahoo.com/evt=51733/*http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
 >


------------------------------------------------------------------------

_______________________________________________
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



--------------------------------------------------------------------------------



_______________________________________________
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



_______________________________________________
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