RE: CE-to-CE PWs, Hierarchical VPLS and Pseudowire Stitching Func tion in draft-stein-pwe3-pwce2e-00.txt

[Sasha Vainshtein <[e-mail]>]

Yaakov and all,
Please see some comments/answers inline.
I have snipped the irrelevant parts of the original text to increase
readability.
----------------------------------------------------------------------------
--------
With best regards,
                          Sasha Vainshtein
email:   [e-mail] <mailto:[e-mail]> 
phone:  [phone] (office)
            [phone] (home)
            [phone] (cellular)

-----Original Message-----
From: Yaakov Stein [mailto:[e-mail]]
Sent: Wednesday, November 12, 2003 3:28 AM
To: Sasha Vainshtein
Cc: Alik Shimelmits; Stewart Bryant (E-mail); Prayson Pate (E-mail); PWE3 WG
(E-mail); David Sinicrope (E-mail)
Subject: RE: CE-to-CE PWs, Hierarchical VPLS and Pseudowire Stitching
Function in draft-stein-pwe3-pwce2e-00.txt




Sasha,

See a few comments below,

[[Sasha]] ...snipped ...

 This decision CANNOT be based on the PW labels in these packets,
because these labels are independently allocated by CE-2 and CE-3
respectively, so that nothing prevents them from being equal
(accidentally or else).
And inspection of the encapsulated data will not help either.

[[Sasha]] ... snipped ... 
[Y(J)S]  You are assuming that these inner labels are
independently downstream allocated by CE2 and CE3,
which doesn't have to be the case.

[Y(J)S] An easy fix is for PW labels to have "structure",
for example (what we do) is to use input-port + output port + serial-number.

[Y(J)S] All that is needed is for the PW labels to be unique in CE1.
[[Sasha]] I do not see this as ever becoming a standard approach.  PW labels
must be understood by the devices that terminate the pWs (in your case -
CE-2 and CE-3).These devices can have their own preferences regarding global
label space etc., and your "structured" labels can easily contradict these
preferences. Not every device is ready to receive and process any value
between 16 and 1M as a valid incoming labels (in fact, most don't). In
addition, you require the labels to be correctly interpreted in two points -
by the CE that terminates them and by the PE that adds transport labels
based on the PW ones. IMO, this is a fundamental contadiction with the MPLS
architecture and hence not acceptable. But let's, at least, present the
complete solution ("structured" non-local labels and all) and discuss it!
The draft did not ever mention "structured" labels (unless I missed
something), and I did not consider an option that clearly contradicts the
basics of the MPLS architecture.

How can this be fixed?

I see three ways to do so.

1. Limit this case to one PW per CE. IMO, this is
   a very problematic limitation (one could say, that
   it is ultimately non-scalable!) and does not
   justify any action wrt the existing documents.

[Y(J)S] Agreed that this is very limited, although a common situation.
[[Sasha]] The problem, as I see it, is that, with your solution, this is a
dead-end situation. The moment your customer requests termination of one
more PW in the same CE, you cannot provide a solution even if the customer
is ready to pay for a more powerful CE!

2. Allow the CE to push both the PW and transport labels
   on top of the payload and PWE3 control info.
   This would make it equivalent to the "normal" PW,and no
   technical changes in the architecture doc are required.
   The transport label can be obtained in many different ways,
   including usage of the MPLS UNI (as suggested by David
   Sinicrope).

[Y(J)S] I don't want to make an MPLS stack here.
[[Sasha]] Franly I do not see it as a problem, but, if it is, let's solve it
in the specific devices that encounter it - this is definitely not a common
problem! 
[[Sasha]] ... snipped ...

 IMO, any technical changes in the architecture document should be
considered
only if we reach a conclusion that advantages of the PW stitching option
outweigh its disadvantages.
  
Hope these notes would be helpful.

[Y(J)S] Very.