Re: Multiple Streams and mutiplex terminations

Hello Javi,

Please see my replies below.

Regards, Christian

On 28/10/2011 9:03 PM, Javi Muñoz wrote:
> Hello Christian,
>
> If we use SDP, an example could be:
>
> MEGACO/1 [123.123.123.1]:11111 (MG <- MGC)
> Transaction = 10 {
> Context = $ {
> Add = ISDNTerm/*BA17-BA18-PRI1*/D/$ {
> Media { Stream = 1
> {
> Local { v=0
> c=ISDN SAPI 16 BA17
> m=data TEI 1 PLP
> c=ISDN SAPI 16 BA17
> m=data TEI 6 PLP
> c=ISDN SAPI 16 BA18
> m=data TEI 3 PLP
> c=ISDN SAPI 16 PRI1
> m=data TEI 9 PLP
> } } },
> Add = HDLCPW1 {
> Media { Stream = 1 {
> # –-Descriptores Local/Remote-- } } }
> } }
>
> Here, I found two serious problems:
[CNG] Attached there's more the above SDP isn't valid. You can't have 
more than one m= line per stream. You can have the "reserve group" 
concept (7.1.7&7.1.8/H.248.1) but the above SDP doesn't match that either.
>
> + We are defining a termination which has ISDN D-channels from 
> different physical interfaces/UNIs (BA17-BA18-PRI1 indicates UNI BRI 
> "17", UNI BRI "18" and UNI PRI "1"). I think it contradicts the 
> definition of terminationss proposed by H.248.1.

[CNG] Yes, so why do you propose a single Termination? Why not add the 
interfaces via multiple Terminations to the Context?

>
> + We are describing several SDP flows (m=) in the Local Descriptor, 
> but these flows are multiplexing in their respective channels D (the 
> situation is not that: aflow is transmitted and, until it ends, no 
> other flow is transmitted; But all flows are simultaneous transmitted, 
> being differentiated by the header of the LAPD frames).
[CNG] OK, I still don't see why you need to specifically label these 
flows via a Stream or SDP m= line.
>
>
> If we use MeGaCo properties to specify the TEI and SAPI, it would have 
> the same problems.
[CNG] Not necessarily.

Below is an example of the use of two properties TEI and SAPI (I haven't 
proposed a descriptor could be LocalControl). By assigning the numbers 
to different termination it indicates which sub-flows go where. i.e. the 
HDLCPW1 termination has flows with TEI=1,3,9 but not 6.

ISDNTerm/BA17
StreamID=1(TEI=[1,6],SAPI=16) \
\
ISDNTerm/BA18 \ HDLCPW1
StreamID=1(TEI=[3],SAPI=16) ---*----- StreamID=1 (TEI=[1,3,9],SAPI=16)
/
ISDNTerm/BA19 /
StreamID=1(TEI=[9],SAPI=16) /




>
> Javi
>
>
> El 28/10/2011 3:23, Christian Groves escribió:
>> Hello Javi,
>>
>> I'm not aware of any SDP that would allow you to specify the TEI and 
>> SAPI. Probably the closest draft to doing this was Tom's draft 
>> http://tools.ietf.org/id/draft-taylor-mmusic-sdp-tdm-01.txt.
>>
>> Rather than use SDP you probably need to create a new package with 
>> two properties that allow you to indicate the SAPI and TEI. Each 
>> property would be a sub-list allowing you to indicate with SAPIs and 
>> TEIs apply to the stream. You probably wouldn't need multiple 
>> StreamIDs in this case.
>>
>> Regards, Christian
>>
>> On 25/10/2011 10:51 PM, Javi Muñoz wrote:
>>> Hello Christian,
>>>
>>> Thank you for your answer.
>>>
>>> I understand MeGaCo syntax doesn't support adding streams in the 
>>> multiplex descriptor neither including N in the multiplex descriptor.
>>>
>>> [Christian] A little, I understand the termination aspect. I'm still 
>>> not sure about the streams. I don't understand how you can have 
>>> multiple H.248 streams for the circuit termination TermA and TermB? 
>>> With what information element (i.e. property) do you assign the SAPI 
>>> and TEI to a Stream? What properties distinguish the Nx64K 
>>> termination from a HDLCW1 one?
>>> [Christian] I'm not 100% sure what you're trying to do but if you're 
>>> only using a single Termination to represent each D-channel and a 
>>> single Termination for the HDLCPW it seems to me that you don't need 
>>> a mux termination and that it may be sufficient to simply use a 
>>> property to assign TEIs to Streams (or Terminations).
>>> [Christian] I think you're trying to solve this at a different level 
>>> than what it should be. The multiplex descriptor describe the 
>>> multiplex of terminations used to make up a stream. It doesn't 
>>> describe what's happening at a stream level itself. We have the 
>>> Stream Descriptor with associated (Local, Remote, LocalControl 
>>> descriptors) to do this.
>>>
>>>
>>> In ISDN, each D-channel (channel of 16 or 64 kbps) transport 
>>> multiple flows muliplexed by its TEI and SAPI in the Q.921 header.
>>>
>>> Every of these flows may need to be transported to a different 
>>> server in the IP network, and each server may receive only a HDLCPW. 
>>> So, I need to multiplex multiple D-channel flows in each HDLCPW. In 
>>> each HDLCPW, each flow would be distinguished by its TEI/SAPI and 
>>> D-channel identifier.
>>>
>>> I can not think how to describe in SDP (Local, Remote, LocalControl 
>>> descriptors) what D-channel flows must be transported by any HDLCPW. 
>>> May you proposed an example?
>>>
>>> Regards,
>>>
>>> Javier Muñoz
>>>>>
>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> Regards,
>>>>>>
>>>>>> Javi Muñoz
>>>>>>
>>>>>>
>>>>>> El 20/10/2011 11:12, Christian Groves escribió:
>>>>>>> Hello Javi,
>>>>>>>
>>>>>>> Did you see my email on the 18/10 which discusses this? Even with
>>>>>>> the clarification of the Terminations there's some questions.
>>>>>>>
>>>>>>> Regards, Christian
>>>>>>>
>>>>>>> On 20/10/2011 7:38 PM, Javi Muñoz wrote:
>>>>>>>> Does MeGaCo support the next contexts?:
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Tmux1=Mux(1x64K(TermA/StreamID=1,TermA/StreamID=2,TermB/StreamID=1)) 
>>>>>>>>
>>>>>>>> -- * -- TermC (HDLCPW1)
>>>>>>>> |
>>>>>>>> TermA (ISDN D-Channel, UNI p) |
>>>>>>>> StreamID=1 (SAPI 16, TEI r)------|
>>>>>>>> StreamID=2 (SAPI 16, TEI s)------|
>>>>>>>> |
>>>>>>>> TermB (ISDN D-Channel, UNI q) |
>>>>>>>> StreamID=1 (SAPI 16, TEI r)------|
>>>>>>>> StreamID=2 (SAPI 16, TEI s)------------|
>>>>>>>> StreamID=3 (SAPI 16, TEI t)------------|
>>>>>>>> |
>>>>>>>>
>>>>>>>> Tmux2=Mux(1x64K(TermB/StreamID=2,TermB/StreamID=3) -- * -- TermD
>>>>>>>> (HDLCPW2)
>>>>>>>>
>>>>>>>>
>>>>>>>> where:
>>>>>>>>
>>>>>>>> (a) Every multiplex termination is multiplexing Streams from
>>>>>>>> different ISDN terminations, and only any Streams of each
>>>>>>>> termination.
>>>>>>>>
>>>>>>>> (b) The result of the multiplexion must be a flow of "64 kbps".
>>>>>>>>
>>>>>>>> Javi Muñoz
>>>>>>>> _______________________________________________
>>>>>>>> Megaco mailing list
>>>>>>>> [email protected]
>>>>>>>> https://www.ietf.org/mailman/listinfo/megaco
>>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Megaco mailing list
>>>>>>> [email protected]
>>>>>>> https://www.ietf.org/mailman/listinfo/megaco
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>
>>> _______________________________________________
>>> Megaco mailing list
>>> [email protected]
>>> https://www.ietf.org/mailman/listinfo/megaco
>>>
>>
>