Re: Ethernet interface in a TDM device: FSC, TDM, or L2SC?

[Dimitri.Papadimitriou@alcatel.be]

hi -


from your drawing, i/f B: TDM and i/f A: L2SC with Ethernet port mode - 
not frame mode - in the port mode case for i/f A the Max = Min LSP 
bandwidth = port capacity 

note: your device is combining the two interfaces i initially describes

the reference i provided allows to link capacities between interfaces (i/f 
A with i/f B) see IACD sub-TLV section

<http://www.ietf.org/internet-drafts/draft-papadimitriou-ccamp-gmpls-mrn-extensions-02.txt>


thanks,
- d.





Freek Dijkstra <fdijkstr@science.uva.nl>
Sent by: owner-ccamp@ops.ietf.org
17/07/2006 12:05
 
        To:     Dimitri PAPADIMITRIOU/BE/ALCATEL@ALCATEL
        cc:     ccamp@ops.ietf.org
        Subject:        Re: Ethernet interface in a TDM device: FSC, TDM, 
or L2SC?


Hi Dimitri,

Thanks for your insights.

> in your case your interface knows about processing payload of the TDM 
LSP, 
> hence the point is if you put such i/f  back to back 
> 
> o) do you switch the STS -> TDM LSP with TDM encoding and TDM ST 
> 
> o) do you switch the whole Ethernet packet flow from port-to-port (no 
> granularity from the Ethernet header info) -> Ethernet LSP with TDM 
> Encoding and L2SC

Neither. The link between two of these kind of interfaces will be pure
Ethernet, no TDM encoding.

Let me make a drawing here.


    Gigabit     +------------+     Gigabit
   Ethernet     |            |     Ethernet in
 ---------------o TDM switch o----------------------
     in UTP   A |            | B   STS-24c in OC-192
                +------------+

My question concerned interface A in the above drawing. It seems that
your descriptions concern interface B instead.

In this case all bytes from interface A are adapted in 24 STS channels,
which are then adapted to the OC-192 of interface B. I can imagine three
lines of reasoning:

* Interface A announces LSC with Ethernet encoding. After all, it does
not look into the headers, and also does no do the TDM encoding (that
happens at interface B)

* Interface A announces TDM ST with Ethernet encoding. After all, it
will embed the Ethernet packets in STS, so it knows about TDM switching.

* Interface A announces L2SC with Ethernet encoding. That seems
incorrect; while the interfaces will probably know about frame
bounderies, it does not process the header.

I assume the first or second is correct.

> ps: if you take such flow from another interface and have limited 
capacity 
> i suggest you take a look at the mln/mln solution doc where you can link 

> such capacities

Do you have a reference?

Regards,
Freek