The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2001-Aug> msg00076



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

draft-ietf-mpls-rsvp-lsp-tunnel-08.txt

  • From: David Charlap <david.charlap@marconi.com>
  • Date: Fri, 10 Aug 2001 13:35:34 -0400

manoj juneja wrote:
> 
> If S1, S2 can be reachable through incoming interface I1 and require
> bandwidth 1M and 2M respectively. The senders S3, S4 falling under
> the same session (as of S1 and S2) require bandwidth 5M and 10M
> respectively but reachable through different incoming interface I2.
> Do I still need to reserve 10M on both the incoming interfaces (i.e.
> I1 and I2) ? Even though we are aware that only 2M (larger of S1,
> S2) will be utilized on I1. Why is it defined like this in the
> protocol ?

Because it's a single _SHARED_ reservation.  It's not four separate
reservations that happen to share resources where convenient.

When the egress router sends an SE-style Resv message, there is exactly
_ONE_ flowspec object transmitted, which covers _ALL_ of the filterspecs
in the flow descriptor.

What I don't understand is why you think you should put LSPs with
different incompatible resource requirements in the same session
together.

-- David