Frank,
I've added the subject line that I forgot on the original posting.
thanks for your reply but I still don't see how it all works. I'm either missing something fundamental or there really is a need for this to be covered in H.323.
I assume that you are referring to 9.3.2/ECMA-333 and 11.2/ECMA-333. In 9.3.2/ECMA-333, it is stated that the Uq-channel number is used as the session identifier and not the other way around, but this is a only a trivial matter.
You have the correct references, but I think a minor mis-interpretation. 9.3.2/ECMA-333 says the pair of logical channels shall be established and then it says that the session numnber shall be used as the Uq-channel number. In other words the session number is determined before the Uq-channel number.
This means that when a QSIG SETUP is received by a gateway that is going to tunnel using the semi-permanent scenario an exchange of H.323 messages must occur to establish the RTP streams and hence the session number before it is possible to forward the QSIG SETUP with the correct Uq-channel number through the H323 tunnel.
The gateway that received the SETUP can only establish the RTP stream in one direction. The RTP stream in the other direction must be initiated by the destination gateway. If that destination gateway is the H.323 slave how does it ensure that it achieves the same session number for the return channel (consider the possibility of multiple QSIG calls arriving at the gateway simultaneously). 11.2/ECMA-333 Note 2 is a reminder that each gateway/PINX must open its own tx channel.
The only way I can think of to ensure both RTP streams have matching session numbers is for the original OLC to be initiated from the slave so that the master can assign the session number and immediately open a return channel with the same session number.
So now the question becomes - what if the QSIG SETUP is originally received at the master gateway. What message does it send to the remote gateway?
Hence my original questions. Or alternatively - what am I missing?
Regards Peter
-----Original Message----- From: frank.derks@philips.com [mailto:frank.derks@philips.com] Sent: 10 February 2003 18:10 To: Peter Price Cc: ITU-SG16@echo.jf.INTEL.COM Subject: Re:
Peter,
I assume that you are referring to 9.3.2/ECMA-333 and 11.2/ECMA-333. In 9.3.2/ECMA-333, it is stated that the Uq-channel number is used as the session identifier and not the other way around, but this is a only a trivial matter.
Your concern seems to be the fact that the "H.245 slave" can establish a QSIG call and therefore tunnel a QSIG SETUP message, but it is not in control over the H.245 session identifier: H.245 states that the session identifier can only be created by the master.
When the slave opens its logical channel towards the master it will specify a session identifier with the value 0. In the OpenLogicalChannelAck message, the master will provide the session identifier that is to be used for this session. If the master overruled the Uq-channel number, it will return this number as the session identifier. If the master accepted the Uq-channel number that was proposed by the slave, it will return this number as the session identifier.
So, unless I don't quite understood you concerns, I do not see any real problems.
Regards,
Frank
Frank Derks |Tel +31 35 6893238 | APPlications |Fax +31 35 6891290 | Philips Business Communications|P.O. Box 32 | |1200 JD Hilversum | |The Netherlands | ----------------------------------------------------| E-mail: mailto:frank.derks@philips.com | WWW: http://www.sopho.philips.com |
Peter Price PeterP@VEGASTREAM.COM Sent by: Mailing list for parties associated with ITU-T Study Group 16 ITU-SG16@echo.jf.INTEL.COM 2003-02-06 11:53 Please respond to Peter Price
To: ITU-SG16@echo.jf.INTEL.COM cc: (bcc: Frank Derks/HVS/BE/PHILIPS) Subject: Classification:All,
ECMA 333 defines the mapping functions for tunneling QSIG over H.323 networks. It describes two scenarios, on-demand and semi-permanent. In on-demand each new ISDN call is tunneled through a separate H.323 call. In semi-permanent multiple ISDN calls can be tunneled through a single H323 call.
10.2.4/H323 V4 describes the mechanism for tunneling ISDN protocols within H323 calls. However, this only ssems to cover the on-demand scenario. It does not include any procedures for the establisment of the additional RTP streams that will be required to support multiple calls.
Obviously the existing logical channel procedures can be used to actually open the channels - thats not the problem.
ECMA 333 states that the Channel ID IE in the tunneled messages should contain the H.323 session id of the logical channel to be used for the audio stream. This isn't a problem for on-demand, there is only one voice channel and it uses the well-known session id 1.
However there are multiple active channels whenn using semi-permanent and the problem is that the session ID is allocated by the H.323 master - what happens if you want to originate a call from the H.323 slave
- what do you
put in the tunneled ISDN SETUP message?
Is this actually covered anywhere or does it need something in H.323? Is there any work in progress within the standards community to cover this issue? Any other commnents?
Thanks
Peter Price Principal Software Engineer Vegastream, a division of Pace Micro Technology plc
Enabling Digital. Our Vision is to be the biggest supplier worldwide of digital gateway technology.
Tel: +44 1344 784914 Fax: +44 1344 784901 Berkshire Court, Western Road, Bracknell, Berkshire. RG12 1RE www.pace.co.uk www.vegastream.co.uk
This E-mail and any attachments hereto are strictly confidential and intended solely for the addressee. If you are not the intended addressee please notify the sender by return and delete the message. You must not disclose, forward or copy this E-mail or attachments to any third party without the prior consent of the sender.
For help on this mail list, send "HELP ITU-SG16" in a message to listserv@lists.intel.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ For help on this mail list, send "HELP ITU-SG16" in a message to listserv@lists.intel.com