Re: Media flow managed by 2 terminations

Thanks Javi for sharing that detailed background!

Indeed, option 3 would be not possible, I quite agree. Option 3 did suppose a single "media flow" component, but not two as in your considered use case.

I could imagine that single-stream-per-termination structures, i.e., dedicated H.248 IP terminations for  PH and FH traffic, might be beneficial, e.g. from perspective of "unsuccessful call handling" like for instance
- H.248.40 for monitoring activity of the media flows on IP side,
- H.248.36 for detection of hanging IP terminations,
- etc.

-Albrecht  

> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of Javi
> Sent: Freitag, 29. Juli 2011 13:46
> To: [email protected]
> Subject: Re: [Megaco] Media flow managed by 2 terminations
> 
> Albrecht,
> 
> The entire network picture is similar to "ETSI_ARGW[TS 183 
> 002 v3.3.1]/Clause 7.3.2" or "draft-bouwen-megaco-isdn-data-00":
> 
> ISDN TE <-> AGW (MG) <-> PH or FH <-> PSPDN
> 
> where PH "Packet Handler" and FH "Frame Handler" are 
> different equipments (I propose a descomposal architecture 
> for PH: "MGC-PH <->
> AGW-PH(SG/MG-PH) <-> PSPDN"; and similar for FH)
> 
> So, AGW/MG terminates ISDN D-channel, and p-,f- LAPD frames 
> would be sent to PH/FH. The IP termination may be: RawFR, 
> IUA, XOT or PWs (HDLCPW, FRPW, TDMPW).
> 
> A D-channel packet mode outgoing call may start with a LAPD 
> p-frame (PLP
> Restart) from ISDN TE, without previous Q.931 signalling. So, 
> a D-channel physical termination is necessary in MG (MG 
> detects that p-frame and sends a MeGaCo notify to MGC; then, 
> MGC sends a SIP INVITE message to PH, signalling the call).
> 
> So, the D-channel physical termination would be neccessary. 
> The doubt would be: two single-stream-per-terminations (an 
> ephemeral termination per packet or frame connection, apart 
> from the D-channel physical
> termination) versus one dual-stream-per-termination (only the 
> D-channel physical termination with two StreamIDs).
> 
> PH and FH are different equipments, and D-channel supports 
> simultaneous packet and frame connections. So, I don't 
> understand alternative "3".
> 
> Javi
> 
> 
> 
> El 28/07/2011 10:43, Schwarz, Albrecht (Albrecht) escribió:
> > Javi,
> > I was hesitating to reply because I was wondering myself, 
> whether I firstly got to know the entire network picture. So 
> far we did consider just a "half-call" model by looking at 
> one termination of the context only.
> > However, I don't guess that the H.248 MG would be the 
> termination point of the D-channel traffic, right?
> >
> > What would be the 2nd termination?
> > Guess this information could clarify the intended function 
> of the H.248 MG, e.g., by just "transparent forwarding" of 
> D-channel frames, or any interworking (?), or ...?
> >
> > That information may provides answers
> > to (3):
> > E.g., the termination is initially configured for "TDM 
> awareness", but "unawareness concerning D-channel traffic 
> type". After detection the termination may become "aware of 
> specific D-channel type frames" ... (but again, would be 
> dependent on your aimed network use case).
> > [Note: sth similar to IP-IP border gateways, initially 
> configured in 
> > media-agnostic mode, possibly later modified to media aware modes, 
> > dependent on ...]
> >
> > to (a,b):
> > The question concerning the alternatives of two 
> > single-stream-per-terminations versus one 
> dual-stream-per-termination 
> > structure could dependent again on your network use case. (E.g., do 
> > you consider a physical or ephemeral termination; any common 
> > properties on termination level?)
> >
> > Just some thoughts, Albrecht
> >
> >
> >
> >    
> >> -----Original Message-----
> >> From: Javi [mailto:[email protected]]
> >> Sent: Mittwoch, 27. Juli 2011 20:02
> >> To:[email protected]
> >> Cc: Schwarz, Albrecht (Albrecht)
> >> Subject: Re: [Megaco] Media flow managed by 2 terminations
> >>
> >> Thanks Albrecht.
> >>
> >> I don't undesrtand your alternative "3)". May you clarify it?
> >>
> >> Also, if the D-channel carries p- or/and f-type frames, what is 
> >> preferable?:
> >>
> >> a) An ephemeral termination for each one (at least, for p- anf 
> >> f-frames).
> >>
> >> b) An streamID for each one (alternative "2").
> >>
> >> Thanks,
> >>
> >> Javi
> >>
> >>
> >>
> >>
> >>
> >> El 15/07/2011 13:24, Schwarz, Albrecht (Albrecht) escribió:
> >>      
> >>>> Does H.248.1 allow this?
> >>>>
> >>>>          
> >>> I think so.
> >>> Guess that following aspect might be also relevant:
> >>>
> >>> Does the D-channel carry
> >>> a) just a single type of frames (e.g. exclusively X.25 
> traffic); or
> >>> b) a multiplex of s-, p- or/and f-type frames (e.g., a mix
> >>>        
> >> of Q.921 signalling, X.25 packet and frame relay traffic).
> >>      
> >>> Possible alternatives to the considered ephemeral
> >>>        
> >> termination for the "X.25/D-channel/{16|64}TDM bearer":
> >>      
> >>> 1) static configuration, i.e., without any H.248 context,
> >>>        
> >> see ETSI TS
> >>      
> >>> 183002, § 7.3.2.2
> >>> ->   using an X.25-over-SCTP bearer
> >>> ->   7.3.2.4.2	IUA/SCTP for X.25 or frame relay traffic
> >>>
> >>> 2) H.248 Mux termination
> >>> use case here: Figure 9 (ex 6)/H.248.1 - Multiplexed termination 
> >>> scenario - Single-to-multiple terminations
> >>> ->   Circuit 1 = D-channel
> >>> ->   StreamID = x: p-type (X.25) frames
> >>>
> >>> 3) SDP for TDM (old proposal: draft-taylor-mmusic-sdp-tdm-01.txt)
> >>> ->   single physical termination
> >>> ->   initially 'TDM'
> >>> ->   after detection modified to 'PKT/TDM'
> >>>
> >>> The attributes
> >>>      a=netprof:<profile>
> >>>      a=e2eprof:<profile>
> >>> could be used to "profile the D-channel" for the specific
> >>>        
> >> type frames
> >>      
> >>> Just some thoughts,
> >>> Albrecht
> >>>
> >>>
> >>>
> >>>
> >>>        
> >>>> -----Original Message-----
> >>>> From:[email protected]
> >>>> [mailto:[email protected]] On Behalf Of Javi
> >>>> Sent: Mittwoch, 13. Juli 2011 17:36 To:[email protected]
> >>>> Subject: [Megaco] Media flow managed by 2 terminations
> >>>>
> >>>> Is it possible to manage a same flow by 2 termination (one
> >>>>          
> >> physical
> >>      
> >>>> and another ephymeral)?
> >>>>
> >>>>
> >>>> Specifically, I need to manage the packet flows (SAPI 16) in a 
> >>>> D-channel ISDN. My proposal is to use:
> >>>>
> >>>> + A physical termination to manage the overall 
> D-channel. When that
> >>>> termination detects a packet flow (SAPI 16/TEI x), it 
> would send a 
> >>>> notify to the MGC.
> >>>>
> >>>> + MGC would create an ephemeral termination to manage that
> >>>> packet flow
> >>>> (SAPI 16/TEI x) and the physical termination should continue to 
> >>>> manage the rest of packet flows (with different TEI) in that 
> >>>> D-channel.
> >>>>
> >>>> Does H.248.1 allow this?
> >>>>
> >>>>
> >>>> _______________________________________________
> >>>> 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
>