Conflicting text in H.323 concerning the requirement for esta blishing a H.245 control channel??

Paul E. Jones paulej at PACKETIZER.COM
Tue Mar 20 16:13:03 EST 2001


Frank, Chris, Charles, et al,

I've been involved in discussions like these many times in the past.  There
are a few things I can recommend, but the most important is: follow what is
stated in H.323v4 :-)

H.323v4 still contains ambiguities, I am quite sure.  However, we tried to
tighten what an endpoint should do.  In particular, things like H.245
tunneling is *strongly* recommended.  We also have a means of doing Fast
Connect and H.245 capability exchange in parallel-- they don't have to
follow each other.  V4 endpoints are now required to set canMapAlias to
true, etc.  Of course, we have to support older behavior for backward
compatibility, but at the current pace, I expect that V2 devices will be
phased out of existence over the next few years: it's always best to look to
the newest version of H.323 to get some idea of the "best current practice".
While H.323 offers a number of options, we tried to make it clearer in some
cases as to what should be done.

I suppose this doesn't necessarily answer any of the questions, but I did
want to try to encourage people to look toward version 4.  Even if you
implement a V2 device, it's wise to consult V4 to make sure that we did not
change behavior in V4 that will introduce backward compatibility concerns.
Theoretically, there shouldn't be any, but I would hate to see somebody
implement H.323v2 *today* and not consider what V4 says about tunneling,
fast connect, mapping aliases, etc.  There are known race conditions in Fast
Connect and H.245.  We put language into H.323v4 as advisory text to prevent
implementers from making the wrong choice.  We carefully crafted that text
to ensure that all known endpoints would operate properly, but it would be
very bad to see new H.323v2 endpoints enter the market that break the rules
spelled out in v4.

In summary, I'd encourage people to read or consult v4-- even if not
implementing v4.  If there are ambiguities in v4, I'll be more than happy to
try to get those clarified.

Paul

----- Original Message -----
From: "Chris Wayman Purvis" <cwp at ISDN-COMMS.CO.UK>
To: <ITU-SG16 at mailbag.cps.INTEL.COM>
Sent: Tuesday, March 20, 2001 12:25 PM
Subject: Re: Conflicting text in H.323 concerning the requirement for esta
blishing a H.245 control channel??


> Charles,
>
> > Part of  establishing a "point-to-point" call involves opening 2 TCP
> > connnections using the Fast Connect procedure as you described it.  If
that
> > is the case, then the extract from H.323v2 below is misleading(I
believe).
> >
> > H.323v2: 8.1.7 Fast Connect Procedure
> >
> > "..... The Fast Connect procedure allows the endpoints to establish a
basic
> > point-to-point call with as few as one round-trip message exchange,
enabling
> > immediate media stream delivery upon call connection."
>
> I don't see a conflict between that quotation and what I said.  You have a
> basic point-to-point call quickly.  Media flow.  Then you set up the H.245
> session with all reasonable speed WHILE MEDIA ARE FLOWING.  So you get
your
> media channels set up as a result of fastConnect, and THEN negotiate other
> capabilities, master-slave etc.
>
> Oh, and there's no need for a second TCP connection anyway - use H.245
> tunnelling.
>
> Regards,
> Chris
>
>
> >> -----Original Message-----
> >> From: Chris Wayman Purvis [mailto:cwp at ISDN-COMMS.CO.UK]
> >> Sent: Tuesday, March 20, 2001 5:47 PM
> >> To: ITU-SG16 at MAILBAG.INTEL.COM
> >> Subject: Re: Conflicting text in H.323 concerning the requirement for
> >> esta blishing a H.245 control channel??
> >>
> >>
> >> Charles,
> >>
> >> It does NOT defeat ANY of the stated aims of FastConnect.
> >> These aims were to get agreed media channels in both
> >> directions open as
> >> quickly as possible.  Doing FastStart AND H.245 gives you your media
> >> quickly, and means you have the power of H.245 thereon.
> >>
> >> In-band DTMF transfer may be used.  If you happen to be using
> >> a codec that
> >> supports it.  If you assume it when you're using an
> >> unsuitable codec you'll
> >> have a problem.  Which is a reason for using H.245 capability
> >> negotiation.
> >>
> >> Regards,
> >> Chris
> >>
> >> Agboh, Charles wrote:
> >>
> >>
> >>> which defeats the whole point of having a Fast Connect
> >>
> >> procedure (FS +
> >>
> >>> H.245).  Why isn't in-band- DTMF transfer used instead (in FS)?
> >>>
> >>> -Charles
> >>>
> >>>
> >>>
> >>>> -----Original Message-----
> >>>> From: Frank Derks [mailto:frank.derks at PHILIPS.COM]
> >>>> Sent: Tuesday, March 20, 2001 4:40 PM
> >>>> To: ITU-SG16 at MAILBAG.INTEL.COM
> >>>> Subject: Re: Conflicting text in H.323 concerning the
> >>>
> >> requirement for
> >>
> >>>> establishing a H.245 control channel??
> >>>>
> >>>>
> >>>> Chris,
> >>>>
> >>>> I thought I was being clear enough, so let me try again.
> >>>> 6.2.8/H.323 states
> >>>> that an enpoint must open one (and exactly one) H.245 control
> >>>> channel. When
> >>>> Fast Connect is being used, I assume that the intention is
> >>>> that no such control
> >>>> channel is opened.
> >>>>
> >>>> To be compliant with 6.2.8/H,323 I would have to open a H.245
> >>>> control channel
> >>>> irrespective of which type of H.245 procedures I will be
> >>>> using. So if I intend
> >>>> to use Fast Start (and assuming that the other party also
> >>>> supports this), I
> >>>> still have to open a H.245 control channel.
> >>>>
> >>>> Frank
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>> cwp at isdn-comms.co.uk on 20-03-2001 15:17:14
> >>>> To:     Frank Derks/HVS/BE/PHILIPS at EMEA2
> >>>> cc:     ITU-SG16 at mailbag.cps.INTEL.COM@SMTP
> >>>> Subject:        Re: Conflicting text in H.323 concerning the
> >>>> requirement for establishing a H.245 control channel??
> >>>> Classification:
> >>>>
> >>>> Frank,
> >>>>
> >>>> Why do you consider this text to be "conflicting"?
> >>>> Specifically, with what does it conflict?
> >>>>
> >>>> Regards,
> >>>> Chris
> >>>>
> >>>>
> >>>>
> >>>>> 6.2.8/H.323 states: "The endpoint shall establish exactly
> >>>>
> >> one H.245
> >>
> >>>>> Control Channel for each call that the endpoint is
> >>>>
> >>>> participating in."
> >>>>
> >>>>
> >>>>> 8.1.7/H.323 never states that when Fast Connect is being
> >>>>
> >> used such a
> >>
> >>>>> control channel should be established. As far as I understand the
> >>>>> mechanism this is only required to switch to "normal" H.245
> >>>>
> >>>> procedures.
> >>>>
> >>>>
> >>>>> It would seem that section 6.2.8 should be rephrased to
> >>>>
> >>>> make clear that
> >>>>
> >>>>
> >>>>> the H.245 control channel shall only be established when
> >>>>
> >>>> "normal" H.245
> >>>>
> >>>>
> >>>>> procedures are being followed and not in the fast connect case.
> >>>>>
> >>>>> Frank
> >>>>
> >>>> --
> >>>> Dr Chris Purvis -- Development Manager
> >>>> ISDN Communications Ltd, The Stable Block, Ronans, Chavey Down Road
> >>>> Winkfield Row, Berkshire.  RG42 6LY  ENGLAND
> >>>> Phone: +44 1344 899 007
> >>>> Fax:   +44 1344 899 001
> >>>>
> >>>>
> >>>
> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >>
> >>>> For help on this mail list, send "HELP ITU-SG16" in a message to
> >>>> listserv at mailbag.intel.com
> >>>>
> >>>
> >>>
> >>>
> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >>
> >>> For help on this mail list, send "HELP ITU-SG16" in a message to
> >>> listserv at mailbag.intel.com
> >>
> >>
> >> --
> >> Dr Chris Purvis -- Development Manager
> >> ISDN Communications Ltd, The Stable Block, Ronans, Chavey Down Road
> >> Winkfield Row, Berkshire.  RG42 6LY  ENGLAND
> >> Phone: +44 1344 899 007
> >> Fax:   +44 1344 899 001
> >>
> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >> For help on this mail list, send "HELP ITU-SG16" in a message to
> >> listserv at mailbag.intel.com
> >>
>
>
> --
> Dr Chris Purvis -- Development Manager
> ISDN Communications Ltd, The Stable Block, Ronans, Chavey Down Road
> Winkfield Row, Berkshire.  RG42 6LY  ENGLAND
> Phone: +44 1344 899 007
> Fax:   +44 1344 899 001
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> For help on this mail list, send "HELP ITU-SG16" in a message to
> listserv at mailbag.intel.com
>

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
For help on this mail list, send "HELP ITU-SG16" in a message to
listserv at mailbag.intel.com



More information about the sg16-avd mailing list