The MPLS WG Archive

["Abes, Andi" <aabes@quarrytech.com>]

Dimmitry,


No, I did mean different Tunnel ID's.

Lets say that there are LSR's Ra Rb Rc and Rd.

The goal we are trying to achieve, if I understand 
you correctly, is to have some LSP's from Ra and Rb
going to Rd share resources among themselves, 
but not with LSP's from Rc going to the same place.

The way I would configure this is:
- Ra and Rb use a session defined by Egress = Rd 
  tunnel = T1
- Rc use a session defined by Egress = Rd and 
  tunnel = T2.

In all sessions, the extended ID is set to 0,
such that if you choose to share its resources
you only need to configure the new ingress to use
the same tunnel ID.

According to your suggestion - how would you configure
these routers ?

Andi.


In any case, it might be good to take this offline
so we don't generate too much traffic on the list.

Andi.




> -----Original Message-----
> From: Dimitry Haskin [mailto:dhaskin@nexabit.com]
> Sent: Thursday, April 13, 2000 5:00 PM
> To: Abes, Andi; mpls@UU.NET
> Subject: RE: FW: I-D
> ACTION:draft-ietf-mpls-rsvp-tunnel-applicability-01.t xt
> 
> 
> > 
> > What is the advantage of doing this,
> > as opposed to simply using different Tunnel ID's
> > and setting the extended tunnel ID to 0 
> > (as the spec suggests) ?
> 
> I assume you meant to write 'the same Tunnel ID'. The 
> advantage is that you
> can confine the session scope to a set of cooperating nodes 
> without risking
> that a node outside of the set can assign the same session 
> object to its
> RSVP session. 
> 
> > 
> > Andi.
> 
> Dimitry
> 
> > 
> > 
> > > -----Original Message-----
> > > From: Dimitry Haskin [mailto:dhaskin@nexabit.com]
> > > Sent: Thursday, April 13, 2000 4:27 PM
> > > To: Abes, Andi; Dimitry Haskin; David Charlap; mpls@UU.NET
> > > Subject: RE: FW: I-D
> > > ACTION:draft-ietf-mpls-rsvp-tunnel-applicability-01.t xt
> > > 
> > > 
> > > > Dimitry,
> > > > 
> > > > Yes, you're right - 
> > > > it will probably work - but it will defeat the purpose 
> > > > of the extended ID.
> > > > 
> > > No, it will not if used judiciously by cooperating nodes.
> > > 
> > > > Andi.
> > > > 
> > > > 
> > > > > -----Original Message-----
> > > > > From: Dimitry Haskin [mailto:dhaskin@nexabit.com]
> > > > > Sent: Thursday, April 13, 2000 4:12 PM
> > > > > To: David Charlap; mpls@UU.NET
> > > > > Cc: Abes, Andi
> > > > > Subject: RE: FW: I-D
> > > > > ACTION:draft-ietf-mpls-rsvp-tunnel-applicability-01.t xt
> > > > > 
> > > > > 
> > > > > A small but not insignificant correction.
> > > > > 
> > > > > > > ...
> > > > > > > 1. For LSP's to be belong to the same session they need
> > > > > > >    to share the same egress point and tunnel ID.
> > > > > > >    If the exteneded tunnel ID is set to the Ingress IP 
> > > > > address, only
> > > > > > >    LSP's originating at the same ingress could ever 
> > > > belong to the
> > > > > > >    same session.
> > > > > > 
> > > > > > Yes.
> > > > > > 
> > > > > 
> > > > > There is nothing to prevent nor it is an error for LSPs 
> > > > originating at
> > > > > different ingress nodes to share the same extended tunnel ID 
> > > > > even if this ID
> > > > > happen to be set to an address of one of the ingress nodes.
> > > > > 
> > > > > Dimitry
> > > > > 
> > > > 
> > > 
> > 
>