Leaving the SIP-XMPP discussion aside (nobody will change their mind anyway) a matter of concern is the 100% dominance of protocol experts in this discussion and no real participation from conference service operators and conference platform vendors. This raises several flags... As a participant in developing IP conference services, I have seen the challenge of integrating IP conference technology with AAA (OSS) and with operator support services. In this light, I believe we need to start with generic requirements for XCON target systems (what should they be anyway?) and examine/compare the candidate protocols later. Such a document SHOULD be co-authored by (1) conferencing services providers first and (2) by conferencing platform vendors, second. Could we shift this debate from protocols to requirements? Henry > -----Original Message----- > From: Rohan Mahy [mailto:rohan@cisco.com] > Sent: Thursday, August 21, 2003 9:55 AM > To: Henry Sinnreich > Cc: 'vinton g. cerf'; 'Marshall Rose'; 'Peterson, Jon'; ietf@ietf.org; > Alan Johnston; Robert Sparks > Subject: Re: WG Review: Centralized Conferencing (xcon) > > Greetings. > > Henry, If we wanted to do a SIP-only conferencing solution, we would > have left this work in SIPPING. However we discovered that several of > the problems we were trying to solve were applicable to non-SIP > signaled conferences (including mixed protocol conferences). We also > wanted to get wider input from the IETF community since there is a lot > of previous work that's gone on. Just like RTP/RTCP streams are setup > by SIP, H.323, MGCP, Megaco, and even proprietary signaling protocols, > there is no reason that a common approach for conference policy, > rosters, inter-conference media manipulation, and floor control can't > be used by vendors which use any signaling protocol. (I will remind > you that, for example, my company has an installed base of over 2 > million phones which are using a proprietary protocol.) Please rest > assured however that SIPPING would make sure that any SIP-specific > pieces get done in SIPPING in a way that works with the output of XCON. > > As for Marshall's comments: there is certainly nothing nefarious about > having a large body of source work when trying to start a Working Group > (I cite XMPP as a recent example). What is important is for folks to > focus on the *charter text* and the desired output, rather than make > assumptions about the intent of those contributors. As for his > suggestion that a generic conferencing working group should live in > APPS instead of TSV, please consider that MMUSIC was the original host > for conferencing work in IETF and therefore there is some precedent for > the work to happen in TSV. I am fine with either, but I believe this > issue was already worked out among Ted, Ned, Allison, and Jon. > Marshall, if there is something specific that you don't like about the > charter as a generic conferencing working group, please propose > alternate rewording. > > thanks, > -rohan > co-chair SIP and SIPPING WGs > > On Wednesday, August 20, 2003, at 06:33 AM, Henry Sinnreich wrote: > > Given the IM background in this discussion, I fail to see why there > > should > > be an IM-originated conferencing option to confuse everyone, when SIP > > conferencing is supporting all types of media and presence and IM just > > as > > well (plus events, user preferences, mobility etc.) in a consistent > > way. > > > > It is high time the IETF should get its act together and converge on > > the one > > single multiparty (conferencing!) multimedia session protocol: SIP. And > > avoid such confusions as Marshall does in the attached. > > > > Look at the huge penetration of SIP with wired and mobile service > > providers, > > as well as on tens of millions of desktops: The IETF SIP work is > > already the > > de facto standard. Let's just stay focused. > > > > It is thus entirely appropriate XCON should be a SIP oriented WG for > > centralized conferencing. > > > > Thanks, > > > > Henry Sinnreich > > MCI > > > >> -----Original Message----- > >> From: vinton g. cerf [mailto:vinton.g.cerf@mci.com] > >> Sent: Wednesday, August 20, 2003 7:39 AM > >> To: Marshall Rose; Peterson, Jon > >> Cc: ietf@ietf.org > >> Subject: Re: WG Review: Centralized Conferencing (xcon) > >> > >> As a prospective supplier of SIP-based services, I am very interested > >> in > >> seeing SIP-based definitions for the support of a wide range of > >> conferencing tools ranging from voice/video to IM and mixtures that > >> might > >> include a participant with only a phone and a fax machine. This is > >> not to > >> say that I would reject other protocol bases for such service but > >> rather > >> to say that we have a significant investment in SIP-based services and > >> would like to see them expanded in standard ways so as to encourage > >> interworking among parties offering such services. > >> > >> I leave it to the IESG and other interested parties to figure out how > >> best > >> to achieve that objective. Perhaps a SIP-oriented WG is the > >> appropriate > >> vehicle, recognizing that what ever procedures are invented, rooted > >> in the > >> SIP system, might well have counterparts in other signalling > >> enviroments > >> and could therefore be re-incarnated in them. Whether that would > >> confer > >> interworking between the SIP and non-SIP systems is beyond my ability > >> to > >> predict. > >> > >> Vint > >> > >> At 03:29 PM 8/19/2003 -0700, Marshall Rose wrote: > >>> jon - sorry for the delay in replying. > >>> > >>> fundamentally, i think it comes down to accuracy in labelling. if > >>> the sip > >>> folks want to do conferencing, then they should have a working group > >>> to > >> do > >>> that. however, the charter for that working group should not imply > >>> that > >> the > >>> scope of the working group is anything beyond sip. > >>> > >>> a reasonable person reading the charter would conclude that the > >>> scope of > >> the > >>> working group is somewhat more generic than sip. > >>> > >>> if the goal for this working group is to be generic, then the > >>> charter is > >>> likely unacceptable since it assumes "facts not entered into > >>> evidence", > >>> i.e., it is sip-centric, and there is a fair body of deployed work > >>> that > >>> manages to do conferencing very well without using that acronym. if > >>> that > >> is > >>> not the intention, then i suggest that the working group be called > >>> something like sipxcon to avoid any confusion. > >>> > >>> as to whether the working group belongs in apps or tsv, a generic > >>> conferencing working group clearly belongs in apps. however, a sip- > >> specific > >>> working group can probably comfortably reside in either. > >>> > >>> /mtr > >> > >> Vint Cerf > >> SVP Architecture & Technology > >> MCI > >> 22001 Loudoun County Parkway, F2-4115 > >> Ashburn, VA 20147 > >> 703 886 1690 (v806 1690) > >> 703 886 0047 fax > >> vinton.g.cerf@mci.com > >> www.mci.com/cerfsup > >