Re: [splices] SIP INVOKE method v1

"Hutton, Andrew" <andrew.hutton@siemens-enterprise.com> Wed, 15 June 2011 09:01 UTC

Return-Path: <andrew.hutton@siemens-enterprise.com>
X-Original-To: splices@ietfa.amsl.com
Delivered-To: splices@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 06E081F0C40 for <splices@ietfa.amsl.com>; Wed, 15 Jun 2011 02:01:45 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level:
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xvryfSWL5MgU for <splices@ietfa.amsl.com>; Wed, 15 Jun 2011 02:01:44 -0700 (PDT)
Received: from mail27.messagelabs.com (mail27.messagelabs.com [193.109.254.147]) by ietfa.amsl.com (Postfix) with SMTP id 4C8ED1F0C35 for <splices@ietf.org>; Wed, 15 Jun 2011 02:01:43 -0700 (PDT)
X-VirusChecked: Checked
X-Env-Sender: andrew.hutton@siemens-enterprise.com
X-Msg-Ref: server-5.tower-27.messagelabs.com!1308128492!34971616!1
X-StarScan-Version: 6.2.17; banners=-,-,-
X-Originating-IP: [62.134.46.10]
Received: (qmail 15919 invoked from network); 15 Jun 2011 09:01:32 -0000
Received: from unknown (HELO senmx12-mx) (62.134.46.10) by server-5.tower-27.messagelabs.com with SMTP; 15 Jun 2011 09:01:32 -0000
Received: from MCHP063A.global-ad.net (unknown [172.29.37.61]) by senmx12-mx (Server) with ESMTP id 0708023F03E6; Wed, 15 Jun 2011 11:01:42 +0200 (CEST)
Received: from MCHP058A.global-ad.net ([172.29.37.57]) by MCHP063A.global-ad.net ([172.29.37.61]) with mapi; Wed, 15 Jun 2011 11:01:42 +0200
From: "Hutton, Andrew" <andrew.hutton@siemens-enterprise.com>
To: "Shekh-Yusef, Rifaat (Rifaat)" <rifatyu@avaya.com>, Paul Kyzivat <pkyzivat@cisco.com>, "splices@ietf.org" <splices@ietf.org>
Date: Wed, 15 Jun 2011 11:01:38 +0200
Thread-Topic: [splices] SIP INVOKE method v1
Thread-Index: AcwnpwHyWdPZ0GtcQdet0Lc+UVQPIgCJvnyQADiYGNAAANdDQAABoVVgAB/pxSA=
Message-ID: <101C6067BEC68246B0C3F6843BCCC1E30C22978813@MCHP058A.global-ad.net>
References: <6369CB70BFD88942B9705AC1E639A33822CD81384C@DC-US1MBEX4.global.avaya.com> <4DF27461.8060204@cisco.com> <101C6067BEC68246B0C3F6843BCCC1E30C22977E6A@MCHP058A.global-ad.net> <6369CB70BFD88942B9705AC1E639A33822CDA812C4@DC-US1MBEX4.global.avaya.com> <101C6067BEC68246B0C3F6843BCCC1E30C229785C2@MCHP058A.global-ad.net> <6369CB70BFD88942B9705AC1E639A33822CDA8144A@DC-US1MBEX4.global.avaya.com>
In-Reply-To: <6369CB70BFD88942B9705AC1E639A33822CDA8144A@DC-US1MBEX4.global.avaya.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach:
X-MS-TNEF-Correlator:
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Subject: Re: [splices] SIP INVOKE method v1
X-BeenThere: splices@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: "Loosely-coupled SIP Devices \(splices\) working group discussion list" <splices.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/splices>, <mailto:splices-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/splices>
List-Post: <mailto:splices@ietf.org>
List-Help: <mailto:splices-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/splices>, <mailto:splices-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Jun 2011 09:01:45 -0000

Hi Rifaat,

I don't understand the argument as the header/body issue makes no difference to what gets standardized and architecturally it is much cleaner to tunnel the non SIP related application stuff in a body that to use a SIP header for the same purpose.

Having said that I still don't think that either mechanism would fall within the SPLICES charter.

Regards
Andy



> -----Original Message-----
> From: Shekh-Yusef, Rifaat (Rifaat) [mailto:rifatyu@avaya.com]
> Sent: 14 June 2011 18:42
> To: Hutton, Andrew; Paul Kyzivat; splices@ietf.org
> Subject: RE: [splices] SIP INVOKE method v1
> 
> Andy,
> 
> > The list of potential actions and associated parameters is huge and
> for this
> > kind of problem I believe that an XML body would be a much better
> solution. I
> > guess I much prefer XML to BNF.
> >
> This actually calls for a header rather than an XML body, because the
> idea is to control what get standardized and what not and clearly
> define the actions that get standardized. With an XML body you actually
> create a tunnel over SIP and you lose control over what get defined.
> 
> Regards,
>  Rifaat
> 
> 
> > -----Original Message-----
> > From: Hutton, Andrew [mailto:andrew.hutton@siemens-enterprise.com]
> > Sent: Tuesday, June 14, 2011 1:18 PM
> > To: Shekh-Yusef, Rifaat (Rifaat); Paul Kyzivat; splices@ietf.org
> > Subject: RE: [splices] SIP INVOKE method v1
> >
> > Hi Rifaat,
> >
> > I am not saying that the current proposal is not extensible but I
> believe that
> > an XML body would be a much better solution to the extensibility
> problem that
> > using a SIP header.
> >
> > The list of potential actions and associated parameters is huge and
> for this
> > kind of problem I believe that an XML body would be a much better
> solution. I
> > guess I much prefer XML to BNF.
> >
> > There is also an architectural issue as the actions don't relate
> directly to
> > the management of the SIP session and therefore most likely need to
> passed to
> > the application layer in the SIP UA to handle. It is much cleaner
> > architecturally if the body can simply be removed and passed to the
> > responsible component in the UA rather than the SIP component having
> to
> > extract the information from SIP headers.
> >
> > Regards
> > Andy
> >
> >
> >
> >
> > > -----Original Message-----
> > > From: Shekh-Yusef, Rifaat (Rifaat) [mailto:rifatyu@avaya.com]
> > > Sent: 14 June 2011 17:32
> > > To: Hutton, Andrew; Paul Kyzivat; splices@ietf.org
> > > Subject: RE: [splices] SIP INVOKE method v1
> > >
> > > Andy,
> > >
> > > > The issue is I believe extensibility which any mechanism for
> > > > invoking services will need and this is the reason why an XML
> body
> > > would be
> > > > better than using SIP headers.
> > >
> > > What makes you think that the current proposal is not extensible?
> > >
> > > Regards,
> > >  Rifaat
> > >
> > >
> > > > -----Original Message-----
> > > > From: splices-bounces@ietf.org [mailto:splices-bounces@ietf.org]
> On
> > > Behalf Of
> > > > Hutton, Andrew
> > > > Sent: Monday, June 13, 2011 10:04 AM
> > > > To: Paul Kyzivat; splices@ietf.org
> > > > Subject: Re: [splices] SIP INVOKE method v1
> > > >
> > > > Hi All,
> > > >
> > > > Whilst I still question whether the mechanism described in this
> draft
> > > actually
> > > > falls within the SPLICES WG charter as it defines new mechanism I
> > > have a few
> > > > comments.
> > > >
> > > > I tend to agree with Paul's comments below but would go further
> to
> > > say that
> > > > even some of the supposedly simply actions such as "decline" &
> > > "ignore" are
> > > > also features which will be difficult to agree the semantics for.
> All
> > > of the
> > > > actions will need to be specified rigorously a quick look at
> ECMA-269
> > > which
> > > > specifies these kind of services for CSTA will give you an idea
> of
> > > what would
> > > > need to be done. I don't believe that even in the CSTA world
> which
> > > defines a
> > > > very large number of services the services "ignore" or "decline"
> were
> > > > specified.
> > > >
> > > > With regard to the discussion on headers vs body the draft
> states.
> > > >
> > > > "There has been some discussion on the list on the question of
> header
> > > field vs
> > > > message body.  This is not the real issue, as either could
> > > conceivably be
> > > > used.  Instead, the main issue is what is being invoked:  is it a
> > > named
> > > > feature/action/event, or is it a script".
> > > >
> > > > I must have missed the discussion about scripting but I don't
> think
> > > this is
> > > > the issue. The issue is I believe extensibility which any
> mechanism
> > > for
> > > > invoking services will need and this is the reason why an XML
> body
> > > would be
> > > > better than using SIP headers.
> > > >
> > > > Regards
> > > > Andy
> > > >
> > > >
> > > > > -----Original Message-----
> > > > > From: splices-bounces@ietf.org [mailto:splices-
> bounces@ietf.org] On
> > > > > Behalf Of Paul Kyzivat
> > > > > Sent: 10 June 2011 20:46
> > > > > To: splices@ietf.org
> > > > > Subject: Re: [splices] SIP INVOKE method v1
> > > > >
> > > > > This is getting better with age!
> > > > >
> > > > > Splitting off the subscription is good.
> > > > >
> > > > > An area that still needs work is in the semantics of the urns.
> > > > > Some of them are more clear than others. Some are "limited" in
> ways
> > > > > that
> > > > > might not matter to some but that could prove a problem in the
> end.
> > > > > Specifically:
> > > > >
> > > > >      o  Answer call       - urn:invoke:call:answer
> > > > >      o  Terminate call    - urn:invoke:call:terminate
> > > > >      o  Decline call      - urn:invoke:call:decline
> > > > >      o  Ignore call       - urn:invoke:call:ignore
> > > > >
> > > > > The above need some elaboration, but I think its just work -
> > > probably
> > > > > anybody who does it will end up with an equivalent result.
> > > > >
> > > > >      o  Send call to VM   - urn:invoke:call:sendvm
> > > > >
> > > > > VM is a "feature", and we haven't standardized features much,
> > > except
> > > > > for
> > > > > bliss. How does one know if the recipient understands VM and
> what
> > > to do
> > > > > to send a call there? What if there are multiple VM options
> (e.g.
> > > what
> > > > > to say, etc.) And what if the call isn't a voice call. (Maybe
> VM
> > > now
> > > > > means voice/video mail. But the "call" could be IM.)
> > > > >
> > > > >      o  Hold call         - urn:invoke:call:hold
> > > > >      o  Unhold call       - urn:invoke:call:unhold
> > > > >      o  Mute call         - urn:invoke:call:mute
> > > > >      o  Unmute call       - urn:invoke:call:unmute
> > > > >
> > > > > Similarly these are all "features" that we all have some
> > > expectations
> > > > > about, but that are not rigorously defined. (But perhaps
> > > Mute/Unmute
> > > > > could be formalized by using the 'transducer' option.
> > > > > (Transducer=none?)
> > > > >
> > > > > My concern with all these "features" is that we are being put
> into
> > > a
> > > > > box
> > > > > where we must either:
> > > > > - define them with sufficient rigor
> > > > > - fall back on "do what I mean", which has proved to be a bad
> idea,
> > > > >    because not everybody has the same notion of what they mean.
> > > > >
> > > > >      o  Conference        - urn:invoke:conference:add
> > > > >                           - urn:invoke:conference:remove
> > > > >
> > > > > This one may be clearer due to the existence of xcon specs.
> > > > >
> > > > > Regarding action parameters:
> > > > >
> > > > >      o  media=audio|video|audvid
> > > > >
> > > > > I'm troubled by audvid. I would rather see "audio,video".
> > > Specifically,
> > > > > what is the namespace here, and what does specifying it mean? I
> > > think
> > > > > the namespace could be: one or more values allowable as the
> <media>
> > > > > parameter in the m-line in SDP.
> > > > >
> > > > > Defining exactly what it means when present is a little harder.
> > > When
> > > > > answering a call that has an offer, I guess it could mean:
> refuse
> > > any
> > > > > m-lines with media type that doesn't match one of these, and
> accept
> > > as
> > > > > many of the remaining as you can.
> > > > >
> > > > > For answering an offerless invite, I guess it could mean: offer
> > > > > whatever
> > > > > media you are capable of offering that match one of these, but
> > > don't
> > > > > offer anything that doesn't match.
> > > > >
> > > > > Is there likely to be a need to get more specific? E.g. about
> what
> > > > > codecs to use? (Hopefully not - that could get very messy.)
> > > > >
> > > > >      o  transducer=speaker|headset
> > > > >
> > > > > My deskphone has three possibilities: speaker|headset|handset.
> > > > > And I'm far from certain this is exhaustive. (The clue wg is
> > > defining a
> > > > > much more complex environment.) This at least needs to be
> > > extensible.
> > > > > We
> > > > > may need some sort of extended model of a UA that includes this
> > > sort of
> > > > > stuff.
> > > > >
> > > > >      o  target=<AOR>
> > > > >      o  direction=recvonly|sendonly|sendrecv
> > > > >      o  abort
> > > > >
> > > > > None of this is intended as objection to the work so far. Its
> just
> > > a
> > > > > start at identifying what else there is to do. The whole notion
> of
> > > > > naming features to invoke causes me the most concern.
> > > > >
> > > > > 	Thanks,
> > > > > 	Paul
> > > > >
> > > > > On 6/10/2011 3:08 PM, Shekh-Yusef, Rifaat (Rifaat) wrote:
> > > > > > Hi,
> > > > > >
> > > > > > We have just submitted a new version of the SIP INVOKE method
> > > draft
> > > > > that
> > > > > > does not include implicit subscription.
> > > > > >
> > > > > > https://datatracker.ietf.org/doc/draft-yusef-splices-invoke/
> > > > > >
> > > > > > We would appreciate it if you review the document and provide
> us
> > > with
> > > > > > your feedback.
> > > > > >
> > > > > > Thanks,
> > > > > >
> > > > > > Rifaat
> > > > > >
> > > > > >
> > > > > >
> > > > > > _______________________________________________
> > > > > > splices mailing list
> > > > > > splices@ietf.org
> > > > > > https://www.ietf.org/mailman/listinfo/splices
> > > > > _______________________________________________
> > > > > splices mailing list
> > > > > splices@ietf.org
> > > > > https://www.ietf.org/mailman/listinfo/splices
> > > > _______________________________________________
> > > > splices mailing list
> > > > splices@ietf.org
> > > > https://www.ietf.org/mailman/listinfo/splices