Re: Multiple Streams and mutiplex terminations

Hell Javi,

Please see ny responses below marked [CNG].

Regards, Christian

On 2/11/2011 8:37 PM, Javi wrote:
> Hello Christian,
>
> http://www.ietf.org/mail-archive/web/megaco/current/msg08470.html

[CNG] OK I thought you meant somewhere other than the Megaco list.

>
> Regards,
>
> Javi
>
>
> El 02/11/2011 0:48, Christian Groves escribió:
>> Hello Javi,
>>
>> Which forum is this being discussed in?
>>
>> Regards,
>> Christian
>>
>> On 1/11/2011 10:36 PM, Javi wrote:
>>> Hello Christian,
>>>
>>> Please see my comments below.
>>>
>>> Regards,  Javi
>>>
>>>>
>>>>
>>>>> b) "ISDNTerm/BA17/[TEI 1]" may be transported on HDLCPW1 but 
>>>>> ""ISDNTerm/BA18/[TEI 1]" may be transported on HDLCPW2. How would 
>>>>> you model this?. You would need to indicate the termination name 
>>>>> in the properties:
>>>>>
>>>>> ISDNTerm/BA17
>>>>> StreamID=1(TEI=[1,6],SAPI=16)\
>>>>>                               \   ---- HDLCPW1
>>>>> ISDNTerm/BA18                  \ /     StreamID=1 
>>>>> (TEI=[1-ISDNTerm/BA17,9-ISDNTerm/BA19],SAPI=16)
>>>>> StreamID=1(TEI=[1],SAPI=16) ----*
>>>>>                                / \
>>>>> ISDNTerm/BA19                 /   ---- HDLCPW2
>>>>> StreamID=1(TEI=[9],SAPI=16)  /         StreamID=1 
>>>>> (TEI=[6-ISDNTerm/BA17,1-ISDNTerm/BA18],SAPI=16)
>>>>>
>>>> [CNG] I suspected we'd end up at this case. It sounds like you're 
>>>> trying to turn a single Context into an entire router to get around 
>>>> the fact that you can't have a single physical termination in 
>>>> multiple contexts. Have a look at H.248.64 to see how we handled 
>>>> that for an IP router. How many more Terminations do you envisage 
>>>> in a Context?
>>>>
>>>> You could add TerminationID to the TEI parameter, but if you're 
>>>> having lots of Termination it may quickly become complicated.
>>>
>>> [Javi] Each MG may have over 50000 interfaces/terminations.
>>>
>>>>>
>>>>>
>>>>> [CNG] Yes, so why do you propose a single Termination? Why not add 
>>>>> the interfaces via multiple Terminations to the Context?
>>>>>
>>>>> [Javi] Yes. Indeed, this woud be my solution to the problem: 
>>>>> Create an ephemeral termination for each flow, including flows of 
>>>>> HDLCPWs:
>>>>>
>>>>> ISDNTerm/BA17/D/SAPI16,TEI1 --- * --- HDLCPW1/BA17/D/SAPI16,TEI1
>>>>>
>>>>> ISDNTerm/BA17/D/SAPI16,TEI6 --- * --- HDLCPW2/BA17/D/SAPI16,TEI6
>>>>>
>>>>> ISDNTerm/BA18/D/SAPI16,TEI1 --- * --- HDLCPW2/BA18/D/SAPI16,TEI1
>>>>>
>>>>> ISDNTerm/BA19/D/SAPI16,TEI9 --- * --- HDLCPW1/BA19/D/SAPI16,TEI9
>>>>>
>>>>>
>>>>> I think this is possible.
>>>> [CNG] A ephemeral Termination for an ISDN termination? These are 
>>>> typically physical terminations.
>>>
>>> [Javi] Really, for each interface I would create:
>>>
>>> + A physical termination maintained in the null context: to notify 
>>> new packet calls to the MGC.
>>>
>>> + An ephemeral termination for each packet flow on the interface.
>>>
>>> ISDNTerm/BA17/D [null context]
>>>
>>> ISDNTerm/BA17/D/SAPI16,TEI1 --- * --- HDLCPW1/BA17/D/SAPI16,TEI1
>>>
>>> ISDNTerm/BA17/D/SAPI16,TEI6 --- * --- HDLCPW2/BA17/D/SAPI16,TEI6
>>>
>>> ISDNTerm/BA18/D [null context]
>>>
>>> ISDNTerm/BA18/D/SAPI16,TEI1 --- * --- HDLCPW2/BA18/D/SAPI16,TEI1
>>>
>>> ISDNTerm/BA19/D [null context]
>>>
>>> ISDNTerm/BA19/D/SAPI16,TEI9 --- * --- HDLCPW1/BA19/D/SAPI16,TEI9
>>>
>>> It was discussed at the forum the possibilty of creating more than 
>>> one termination for each physical interface, and concluded that it 
>>> was possible.
>>>
[CNG] I think there is only ever one physical termination. There can be 
empheral terminations to describe the connection topology. Something 
like the above would be possible although you would need to work on your 
naming conventions to ensure they are valid.
>>
>
>