The MPLS WG Archive

[Curtis Villamizar <curtis@workhorse.fictitious.org>]

In message <200311111510.hABFAaxg011064@rtp-core-1.cisco.com>, George Swallow w
rites:
> Curtis -
> 
> > I'm personally not all that convinced about P-MP but I'm also not
> > ready to say it should be eliminated based on it not fitting into a
> > particular OAM framework.
> 
> You need to re-read what Neil wrote.  
> 
> > Loa....you are misrepresenting my 'agreement on the situation'.   I =
> > agree there is a problem.....but has it ever occurred to you that the =
> > starting point may be wrong?  The problem is not that Dave's paper fails =
> > to line-up with the MPLS arch papers, the problem is that the MPLS arch =
> > (sic) stuff is itself architecturally broken......the mere act of =
> > 'doing' mp2p and PHP does not bless them with arch validity please note. =
>           ^^^^
> >  You can't expect Dave to fix that.....no doubt you also saw Dave's =
> > response.
> 
> He's talking about LDP/Unicast.  Which we use for everything except if
> you *only* use TE in the core.  So his view of the world is much more
> narrow than you imagined!
> 
> ...George


Right..  Sorry.  Amend my statement to "I'm also not ready to say we
should eliminate mp2p based on it not fitting into a particular OAM
framework".

LDP/unicast seems here to stay as well.

Curtis