The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2000-Sep> msg00504



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

MPLS/BGP routing question

  • From: Eric Rosen <erosen@cisco.com>
  • Date: Fri, 29 Sep 2000 11:23:39 -0400
  • cc: "'Chris Flores'" <chris.flores@onfiber.com>, "'Javier Antich'" <javier.antich@telindus.es>, "'Michel Redondo Ferrero'" <mredondo@idecnet.com>, mpls@UU.NET
  • User-Agent: EMH/1.10.0 WEMI/1.13.2 (Mochimune) FLIM/1.12.1(Nishinokyō) Emacs/20.6 (sparc-sun-solaris2.5.1)MULE/4.0 (HANANOEN)


Brijesh> given the performance difference in access routers and core routers,
Brijesh> and the role of a route reflector in applying policies on behalf of
Brijesh> clients, it is pretty obvious that a core router is more suitable for
Brijesh> route reflection function. 

I  think this  is a  bit like  comparing apples  and oranges.   The powerful
packet forwarding  engines that one finds  in core routers don't  do the BGP
protocol  processing work.   The  performance difference  in  edge vs.  core
routers is  pretty irrelevant  in this  context.  In any  event, a  big edge
router with lots  of "high touch" features is going to  be a pretty powerful
beast itself. 

But more  to the  point --  when people  ask "do I  need to  run BGP  in the
core?", they aren't asking what kind  of box makes the best route reflector,
they are  asking whether their  core routers (i.e.,  the ones that  are kept
quite busy moving packets between edges) need to run BGP.