The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2003-Feb> msg00082



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

E2E VoIP over MPLS ('VoMPLS') Header Compression - End-to-End MPLS

  • From: Curtis Villamizar <curtis@fictitious.org>
  • Date: Thu, 20 Feb 2003 13:12:18 -0500
  • cc: IETF MPLS List <mpls@UU.NET>


In message <3E550FB2.3060900@marconi.com>, David Charlap writes:
> Curtis Villamizar wrote:
> > 
> > You can't possibly do MPLS RSVP/TE from the customer edge all the way
> > to the other customer edge, so voice/MPLS is an immediate non-starter.
> 
> If used in the context of voice-through-internet, then yes.

You are completely igoring the context of this statement which was if
you try to solve the problem CE to CE then voice/MPLS is a
non-starter.

So I guesss we agree.

> But it may still be useful within a voice carrier.  Think of how voice 
> carriers currently run their voice circuits over ATM.  You don't have an 
> ATM circuit going all the way to your home.  If carriers want to 
> eventually merge their voice and data infrastructures, then voice-over- 
> MPLS sounds like something they may want.

So we're back to ATM again (why doesn't that surprise me), and
specifically ATM used in the metros.  :-)

f so, the CE-PE-...-PE-CE described in the draft just applies within a
metro and so CE-PE-...-PE-CE is really incorrect.  It might just be
that the applicability is seriously overstated in the draft.

> But I don't think there's any need for the IETF to get involved in such 
> development, since the only IP-related component would be managing the 
> LSP itself, and existing protocols are adequate for that task.
> 
> -- David

What we are gradually establishing is that the practical scope of this
work is much more limited than the internet-draft would imply.  Given
the state scope of the internet-draft, the technique it is presenting
is inpractical at best if not infeasible.  It would seem from this
thread that for certain problems, like CE-PE link, and PE to PE across
an entire SP (rather than just within a metro), there are better
solutions.

If this is a technology that for practical reasons is applicable only
for the metro but would in fact be useful for the metro, just have the
internet-draft say that and move on either accepting it acknowledging
its limited applicability or not accepting it.

Curtis