Chris,
There are applications where an IWF can register an EP from one domain into another. This allows automatic visibility of EP from one domain from another. In this case the IWF is registering not only itself but other EPs. For this scenario, the third-party entity is the IWF, right?
regards,
charles
-----Original Message----- From: Chris Wayman Purvis [mailto:cwp@isdn-comms.co.uk] Sent: Friday, November 24, 2000 6:44 PM To: Agboh, Charles Cc: ITU-SG16@mailbag.cps.intel.com Subject: Re: Third party registration/group registration
Charles,
If your application is a gateway, this is NOT third-party registration. It is simply registration of a potentially large number of aliases. The gateway is not registering on behalf of several H.323 endpoints - it is the edge of the H.323 network as far as signalling is concerned.
This appears to be possible in version 4 (although probably not before) of H.225.0 using the "additiveRegistration" field in RRQ.
Regards, Chris
"Agboh, Charles" wrote:
Chris,
Q1: How can I support third party registration in H.323 v2?
A:1 alternateEndpoint structure but I have the UDP packet size limitation so if I want to do third party registration on behalf of thousands of EPs what can I do?
Q2: Is there a clean way to circumvent this limit?
The application is an H.323-X gateway. X may be SIP.
regards,
charles
-----Original Message----- From: Chris Wayman Purvis [mailto:cwp@isdn-comms.co.uk] Sent: Friday, November 24, 2000 5:48 PM To: Agboh, Charles Cc: ITU-SG16@mailbag.cps.intel.com Subject: Re: Third party registration/group registration
Charles,
The way you describe this I don't understand how it is "third-party".
It is possible to register lots of aliases simultaneously to the same endpoint. This is not third-party as I understand the term. The size of
a
single RRQ is limited by your transport network, but if you're using UDP that gives you about 64 kilobytes - is that really not enough for your
purposes?
Maybe you should describe your application, so we can think about how to achieve the desired result.
Regards, Chris
"Agboh, Charles" wrote:
Chris, all;
Thanks for the response. The scenario I am thinking off is this.
-(a) for a single endpointIdentifier -(b) for a single callSignalingAddress/rasAddress -(c) potentially thousands of terminalAlias's.
In version 2, (a) and (b) are not supported together because of the following text in H.225.0 V2:
(d) Section 7.2.2
"If the Gatekeeper receives an RRQ having the same TransportAddress as a previous RRQ and a different alias address, it should
replace
the translation table entries."
My question is, how can I have (a), (b), (c) or third party registration with (c).
I was thinking that the alternateEndpoint Structure may be usful for (c)
but
there is the limitation of the size of the UDP packet. An alternative
could
be to use the lightweight RRQ but from section 7.9.1 of H.225.0:
"An endpoint can send a lightweight RRQ consisting of only keepAlive, endpointIdentifier, gatekeeperIdentifier, tokens, and timeToLive."
The last resort for (a), (b) and (c).
-H.323v4 addtitive registration -Group registration (i.e. e164 prefix registration). This is supported
in
v2->v4. -Is there another method I can use with an H.323 complaint GK without creating backward compatibility problems.
Best regards,
charles
-----Original Message----- From: Chris Wayman Purvis [mailto:cwp@ISDN-COMMS.CO.UK] Sent: Friday, November 24, 2000 2:41 PM To: ITU-SG16@MAILBAG.INTEL.COM Subject: Re: Third party registration
Charles, All,
Is third party registration supported in H.323 V2, V3, V4?
What's your application?
According to "original intent", the answer is "yes". However this will surprise a lot of people and I doubt you'll find many gatekeepers to
support
it. It all comes down to a couple of strange-looking "sequence of"s in
RRQ.
I was reliably informed some time ago that the original intent was that IF
RRQ
gives more than one callSignalAddresses, it should give the same number
of
rasAddresses and either none or the same number of terminalAliases. Similarly if giving more than one rasAddress there should be the same number of callSignalAddresses Then the first element in each list maps to the
first
element of the others, the second to the second etc. My assumption is
that
RCF or RRJ ought to be sent only to the first rasAddress in the list. However, as I said before, this is an undocumented "original intent",
and
the ASN.1 that came out of it is far from the best way to achieve it. I
don't
know if anybody actually handles multiple entries in these fields (apart from terminalAlias) or, if they do, HOW they handle them.
Does anyone out there with a gatekeeper have any input on whether or how they handle this?
Charles,
I'm cross-posting this to the H.323 implementors list - you may get more idea of what's actually implemented from there.
Regards, Chris -- 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@mailbag.intel.comFor 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
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
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ For help on this mail list, send "HELP ITU-SG16" in a message to listserv@mailbag.intel.com