The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2002-Jun> msg00081



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

Basic LDP question...

  • From: "Chetan Pinto" <chetanpinto@hotmail.com>
  • Date: Sun, 16 Jun 2002 13:43:02 -0700
  • Cc: prashant_sharma@trillium.com, mpls@UU.NET, eric.gray@sandburst.com
  • X-OriginalArrivalTime: 16 Jun 2002 20:43:02.0987 (UTC) FILETIME=[68E3A9B0:01C21576]
  • X-Originating-IP: [12.234.67.129]

>Hi,
>   I have a point regarding the Proxy-Egress case. Just the point that you
>don't have a session with the next hop peer shouldn't make you behave as a
>proxy egress. You should also have a knowledge of your switch
>capabilities (may be by means of some configuration), whether it can do
>the IP forwarding or not. Also it may be necessary to know apriori for
>what prefixes I can act as proxy-egress.

Agreed that Proxy egressing should be configurable.
But why does the question "if the switch is capable of IP forwarding"
comes into picture. Isn't it a paradox? If the switch is sending
routing updates to the upstream routing peers , then it needs to take
care of the downstream traffic. Am I missing something(s) here?

>This becomes more significant an
>issue in case of CR-LDP where just like that you can't become a proxy
>egress and terminate the Constrained paths which may not satisfy the
>required service guaranties.

Implicit behaviour of CR-LSPs, but vanilla LSPs shouldn't have an issue

Thanks
chetan

>
>regards
>Bhushana
>
>
>On Fri, 14 Jun 2002, Chetan Pinto wrote:
>
> > >Hi All,
> > >
> > >     I have one basic question in handling of label request message. 
>LRq.2
> > >states that "Is there next hop for FEC ?". My question is what should 
>be
> > >the
> > >LSR behavior when there is next hop for FEC (as per routing module) but
> > >there is NO LDP session with the next hop node.
> > >
> > >    In such cases, should we consider ourself as EGRESS node and send 
>Label
> > >Mapping to upstream node (for both independent and ordered case)?
> >
> > Yeah, the node should act as a Proxy-Egress
> >
> > >Or we
> > >should interpret this as a failure of LRq.2 and send Notification to 
>the
> > >upstream node (again for both independent and ordered case) ?
> >
> > No
> >
> > >   Similarly, what should be LDP behavior when recognize new FEC 
>procedure
> > >is
> > >executed and there is NO LDP session with next hop node.
> > >Should we consider ourself as EGRESS node (till the time session with 
>next
> > >hop is up)? Or we should consider ourself as "intermediate" node 
>always, no
> > >matter session with next hop is up or down?
> >
> > Again act as a proxy-egress
> >
> >
> > >
> > >   Thanks in advance.
> > >
> > >Regards
> > >Prashant.
> > >
> >
> >
> > _________________________________________________________________
> > Join the world’s largest e-mail service with MSN Hotmail.
> > http://www.hotmail.com
> >
> >


############################################################
Chetan Francis Pinto
Phone : Res : 408-244-1352
############################################################

_________________________________________________________________
MSN Photos is the easiest way to share and print your photos: 
http://photos.msn.com/support/worldwide.aspx