Paul,
I'll be happy to present anything. Can't say that I will defend anything, but I'll present it :-)
Past experience shows that on complex issues this simply is not sufficient - there will be too many people who don't understand, hence fear, hence (if it ever comes to a vote) oppose any proposal in either camp. So, realistically, nothing will result. Not that I don't understand that you have plenty else to do at these meetings!
I missed much of the discussion, so if you wouldn't mind, will you write an e-mail me summarizing the two opposing opinions?
There are a three questions I can identify in this discussion. Note that the answers I give with initials CWP by them are mine, PL are Paul Long's views as I understand them, CA are Charles Agboh's views as I understand them. If I have misunderstood the views of either of these people I apologise and would be obliged if they would clarify their positions. I am not intending to misrepresent anybody, but I may have inadvertently done so!
Q1. Is a device permitted to call itself H.323 if it does not support RAS (or can be configured not to use RAS)? A1a (CWP). No, on the grounds that certain RAS messages are mandatory for devices to support (H.225.0 section 7.7) A1b (CA, PL). Yes, on the grounds that gatekeepers are an optional part of an H.323 system.
Assume RAS is being supported in a particular "system" (without prejudging the answer to Q1), and we come to the other question: Q2. If an endpoint has attempted gatekeeper discover and/or registration, and been rejected (GRJ or RRJ) by all gatekeepers it has found (this being at least one gatekeeper), is the endpoint permitted to make and receive H.323 calls? A2a (CWP). No, on the grounds that there is a gatekeeper in control of this system, and it has denied access. Also on grounds of common sense (ARJ makes no sense if an endpoint may deregister and then make the call it's been refused). A2b (PL). Yes, presumably on grounds of restrictiveness.
Q3. Should third-party registration be standardised? A3a (CA). Yes. A3b (CWP). Possibly, but before this can be answered a LOT of clarification is required on what is meant by the term "third-party".
Regards, Chris
----- Original Message ----- From: "Chris Wayman Purvis" cwp@isdn-comms.co.uk To: plong@ipdialog.com; ITU-SG16@mailbag.cps.intel.com; "Paul Jones" paul.jones@ties.itu.ch; rkbowen@cisco.com Sent: Friday, December 01, 2000 11:57 AM Subject: Re: Third party registration/group registration
All,
I think on the matter of whether the standards permit H.323 entities to continue to operate unregistered in an environment where there is a
gatekeeper,
in two cases:
- They have discovered the gatekeeper, but have been rejected by either
GRJ or
RRJ. 2. No attempt has been made to discover a gatekeeper.
I have a view. Paul Long has a view. I have a different view. These
differ.
I believe, however, that whichever view prevails, one must, and this must
be
through the standards themselves. Common usage can not decide this, as
it's a
question of whether something is permitted by the standard. I could draw
up a
very quick proposal (<= 1/2 page) for the next ITU meeting, but I will not
be
able to come and present it. To help the experts at the meeting to come
to
the right decision, however, opposing proposals probably ought to be
presented
properly by their authors, giving the two viewpoints. Any volunteers
(Rick?
Paul J? As editors of the relevant standards?)?
Regards, Chris
Paul Long wrote:
Chris,
Re Re Q2: Yes, I agree with A2c, but I see no point in wasting _any_ bandwidth for RAS on a system that does not contain a gatekeeper.
Re Re Q3: I think we understand each other and that we'll just have to
agree
to disagree on this. It's up to the implementation and ultimately the
market
to decide whether the user may disable RAS.
Re Re Q4: Correct, if something behaves "as if" it were compliant then I suppose it is, well... compliant. :-) No change is necessary to the Recommendations.
Paul Long ipDialog, Inc.
-----Original Message----- From: Chris Wayman Purvis [mailto:cwp@ISDN-COMMS.CO.UK] Sent: Friday, December 01, 2000 8:41 AM To: ITU-SG16@MAILBAG.INTEL.COM Subject: Re: Third party registration/group registration
Paul,
Gettin' closer.
I think (and hope!) so!
Re Q2: I simply can't justify making the user wait several seconds for
a
discovery that will always fail in a system without a gatekeeper
before he
or she can place or answer each and every call. Can you? Therefore,
the
user
should be able to turn off RAS.
So you'll agree on answer c then? What about regular reattempts to find a gatekeeper (excluding my
suggestion
of when a call is attempted).
Re Q3: I agree with you, except that with some endpoints the user may
then
turn off RAS and place or answer calls without RAS. Note that the
typical
user will most likely not do this, since at least placing a call
without a
gatekeeper would require more knowledge than the average user
posseses,
e.g., the IP address of the called party.
This is where I disagree, on the grounds that if we allow calls in
systems
with gatekeepers from endpoints that are not registered, we may as well throw away the gatekeeper altogether (or at least call it a proxy and start
speaking
SIP).
Re Q4: Maybe he has decomposed his endpoint. In the C Standard, there
is
something called the "as if" rule. Applying it here, if the system experiences consistent behavior from a possibly decomposed entity that
is
acting "as if" it were a corporate entity, it is compliant IMO. Who
cares
where messages originate as long as the effect is the same? In a
different
way, the "as if" rule is what allows routing gatekeepers to do what
they
do--they can fiddle with messages streams all they want as long as
they
maintain consistency "as if" the message streams were originating from
a
compliant endpoint.
Surely if the "as if" rule applies there can be no requirement for any changes to the standard - otherwise it isn't "as if"! I grant the possibility of the decomposed endpoint, although I don't personally understand why anyone would want to, as the communication between the decomposed parts would be at least as complicated as RAS itself.
Note that when I say, "user," I mean either the actual user of the
endpoint
or possibly an administrator of the system. I think it's perfectly reasonable to make the use of RAS an administrated setting.
Agreed.
Regards, Chris
For help on this mail list, send "HELP ITU-SG16" in a message to listserv@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