Re: TD-50/Osaka - Changes to the conferenceGoal field

20 Jun 2000

      Pete,
As you well know, for H.323 mandatory is in the eyes of the implementors.
If an implementor does not think that they are useful for him then it won't
be implemented.  I think that Packages can be a useful service, but not
mandatory.  (I have yet to see a specific service to use Packages; but
theoretically they are useful.)
The problem is service that are mandatory, but have optional coding.  Annex
L has two coding formats in order to use v2 clients.  I can seen some
Packages services using the same technique - v4 uses Packages, v2 uses
nonStandardControl.  I would prefer a negotiation process like that for RAS
negotiation - the originators asks for what is wanted, the response is what
is available.  Then the originators chooses what to do.  If this is done,
there is no need to mess with conferenceGoals.
Bob
------------------------------------------------------------------
Robert Callaghan
Siemens Enterprise Networks
Tel: +1.561.923.1756    Fax: +1.561.923.1403
Email:  Robert.Callaghan@ICN.Siemens.com
------------------------------------------------------------------
-----Original Message-----
From: Pete Cordell [mailto:pete@TECH-KNOW-WARE.COM]
Sent: Tuesday, June 20, 2000 2:01 PM
To: ITU-SG16@mailbag.cps.intel.com
Subject: Re: TD-50/Osaka - Changes to the conferenceGoal field
Bob,
You have made the perfect argument for leaving neededFeatures in the
conferenceGoal!
Previously we could say that support for packages was optional because we
were relying on the conferenceGoal backwards compatibility method to work
for V4 endpoints.  This is because all the support packages need is to
reject a call if neededFeatures is signalled.
If we have to move neededFeatures to some other position, then packages MUST
become MANDATORY, at least in terms of recognising that neededFeatures are
present.
I would imagine that you wouldn't like this.
Pete.
P.S. I believe you have expressed some concern that calls should be able to
continue even if certain features are not supported which may be causing you
to be worried about packages.  This can be easily achieved with packages.
Simply DON'T signal the features that you desire as neededFeatures, signal
them as desiredFeatures.  You then have the behaviour that you are looking
for.  The important thing here is that it gives the network architect the
choice of the behaviour that should happen.  Or should Siemens customers
take what they're given and lump it :)
=============================================
Pete Cordell
Tech-Know-Ware
pete@tech-know-ware.com
+44 1473 635863
=============================================
----- Original Message -----
From: Callaghan, Robert Robert.Callaghan@ICN.SIEMENS.COM
To: ITU-SG16@mailbag.cps.intel.com
Sent: 20 June 2000 14:53
Subject: Re: TD-50/Osaka - Changes to the conferenceGoal field
...
Rich,
I agree that a v4 client that supports packages and further supports a
specific feature will respond.  My comment is in regard to the generation
of
...
a negative message indicating that the feature is not support.  If a v4
client supports packages but not a feature, it will resond with a negative
message.  However if the v4 client does not support packages, it will not
respond at all.  This is worse that the v2 or v3 clients if the change in
conferceGoals is made.  Therefor, my idea is that the change in
conferenceGoal is not needed because than the v2, v3, and some v4 clients
will work the same.  This allows the addition of coding of package
information to be anywhere other than conferenceGoal.
Bob
-----Original Message-----
From: Rich Bowen [mailto:rkbowen@cisco.com]
Sent: Tuesday, June 20, 2000 3:20 AM
To: Callaghan, Robert
Cc: 'Mailing list for parties associated with ITU-T Study Group 16'
Subject: Re: TD-50/Osaka - Changes to the conferenceGoal field
Bob,
If the v4 client supports the features listed in the needFeatures field,
then it will respond with those same features listed in a
supportedFeatures field, as an acknowledgement.  Presumably if it
acknowledges features in this way, then it will not ignore packages
related to those features.
Rich
"Callaghan, Robert" wrote:
...
Rich,
I question that the concept for forcing conferenceGoal to fail will even
work.
The use of conferenceGoal will detect a v2 or v3 client through a
failure.
...
...
But the version number will also detect the older versions.  A v4 client
can
...
properly detect and decode the new structure.  However, support of
packages
...
is optional, so the client can ignore the requirement to evaluate the
packages.
Then what?
Bob
-----Original Message-----
From: Rich Bowen [mailto:rkbowen@CISCO.COM]
Sent: Sunday, June 18, 2000 5:14 PM
To: ITU-SG16@mailbag.cps.intel.com
Subject: TD-50/Osaka - Changes to the conferenceGoal field
All,
In Osaka we agreed to add support for packages to H.225.0 along the
lines of TD-50 (ftp://standard.pictel.com/avc-site/0005_Osa/TD-50.zip).
There was some concern raised about the proposed modifications to the
conferenceGoal field in the Setup message.  The meeting report states:
"Some issues were raised about conflicts in the usage of the conference
goal field between this new method and H.450.x.  A possible solution is
to create a new field in the ASN.1 to avoid conflicts.   The editors are
empowered to work with interested parties to resolve these conflicts in
the ASN.1 before the white paper is issued."
The modification proposed in TD-50 was the addition of the
"neededFeatures" field to the confereceGoal field of the Setup-UUIE:
    conferenceGoal          CHOICE
    {
            create          NULL,
            join            NULL,
            invite          NULL,
            ...,
            capability-negotiation          NULL,
            callIndependentSupplementaryService     NULL,

-->             neededFeatures          NeededFeatureGoal
        },
where the NeededFeatureGoal structure is defined as:
NeededFeatureGoal ::= SEQUENCE
{
        basicGoal       CHOICE
        {
                create                                  NULL,
                join                                    NULL,
                invite                                  NULL,
                capability-negotiation                  NULL,
                callIndependentSupplementaryService     NULL,
                ...
        } OPTIONAL,
        features        SEQUENCE OF SupportedFeatures,
        ...
}
An alternative approach would be to add a neededFeatures field at the
highest level of the Setup-UUIE ASN.1 instead of inside the
conferenceGoal structure, similar to the way the desirededFeatures and
supportedFeatures fields will be added (see TD-50), like this:
Setup-UUIE ::= SEQUENCE
{
        [snip]
-->     neededFeatures          SEQUENCE OF SupportedFeatures OPTIONAL,
        desiredFeatures         SEQUENCE OF SupportedFeatures OPTIONAL,
        supportedFeatures       SEQUENCE OF SupportedFeatures OPTIONAL
}
The motivation for adding neededFeatures to the conferenceGoal field was
to force a call failure when trying to setup a call to pre-v4 endpoint
and there is some v4 or later feature that is *required* for the call.
The mechanism is intended to work like this:
o The v4 EP sends Setup to the pre-v4 endpoint, and specifies some
required feature in the neededFeatures field of the conferenceGoal.
o The pre-v4 EP doesn't recognize the neededFeatures field as a
supported CHOICE for a conferenceGoal, so it sends Release Complete.
There would not be any conflicts with H.450, as suggested in the meeting
report, because if the callIndependentSupplementaryService was needed,
it would still be specified.  The difference is that, if a neededFeature
was also required, the H.450 goals would be specified inside the
NeededFeatureGoal structure, which v4 and later endpoints would be aware
of.
These are the pros and cons of adding neededFeatures to the
conferenceGoal vs. adding it to Setup-UUIE, IMO:
o Advantages:
    - *May* force an early call release if a required feature is
      not supported by a pre-v4 endpoint.
o Disadvantages:
    - ASN.1 and thus implementation would be more complex.
      Potentially have to check the conferenceGoal in two structures
      instead of one.
    - If pre-v4 EP sends Release Complete, there is no way to
      know whether it was sent because of an unrecognized
      conferenceGoal, because there is no ReleaseCompleteReason
      defined for that purpose.
    - H.323 does not require failing a call if the conferenceGoal
      is unrecognized, so the mechanism may not work at all.
Although I think this mechanism is a good idea (and I supported it in
Osaka), at this point I don't think it would work reliably enough to
justify the added complexity.  So I would prefer the alternative
approach described above, adding the neededFeatures field directly to
the Setup-UUIE.
Okay, fire away. :-)
Regards,
Rich

Richard K. Bowen                     Cisco Systems, Inc.
VoIP Session Protocols               Research Triangle Park, NC, USA

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

--
Richard K. Bowen                     Cisco Systems, Inc.
VoIP Session Protocols               Research Triangle Park, NC, USA

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

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