The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2003-Mar> msg00366



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

[PWE3] MPLS PID

  • From: "Ferrell, William" <William.Ferrell@titan.com>
  • Date: Thu, 27 Mar 2003 15:17:05 -0500
  • Cc: "'David Allan'" <dallan@nortelnetworks.com>, jeremy.de_clercq@alcatel.be, "'Scott W Brim'" <sbrim@cisco.com>, pwe3@ietf.org, mpls@UU.NET

I see Eric

It seems as though your perspective is more specific to your
implementation.right ?
Will

-----Original Message-----
From: Eric Rosen [mailto:erosen@cisco.com]
Sent: Thursday, March 27, 2003 3:03 PM
To: Ferrell, William
Cc: 'David Allan'; jeremy.de_clercq@alcatel.be; 'Scott W Brim';
pwe3@ietf.org; mpls@UU.NET
Subject: Re: [PWE3] MPLS PID 



Eric> 6. No one has given a  reason why the first nibble of any such control
Eric>    word should not be as Stewart has proposed. 

David> Except that it breaks some  implementations that may have 0x04 in the
David> first nibble without being IP packets. 

Referring to implementations which don't use a martini control word? 

William> clarify  why  the exception  in  #6 is  so  easily  ignored in
your
William> opinion. 

Well, my company  does have implementations which omit  the control word, so
that's not it ;-)

We're in  a situation now in  which we have  things that don't work  so well
together (core routers doing ECMP  with MPLS, edge routers doing pseudowires
without a  martini control word).  If  we want everything  to work together,
something has to give, and it's not going to be something in the core ;-)