The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2002-Jun> msg00154



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

your mail

  • From: Eric Osborne <eosborne@cisco.com>
  • Date: Thu, 27 Jun 2002 14:53:29 -0400
  • Cc: eosborne@cisco.com, scullptor@yahoo.com, mpls@UU.NET
  • User-Agent: Mutt/1.4i
  • X-GPG-Fingerprint: 6412 0836 E440 B3EA 980C 4951 611E 1819 2E71 8562

On Thu, Jun 27, 2002 at 05:03:25PM +0000, George Sheng wrote:
> ] On Thu, Jun 27, 2002 at 07:50:39AM -0700, Richard Stephens wrote:
> ] > Is there any automated mechanism in existing MPLS
> ] > drafts that enable the automated generation of two
> ] > lsp's given two equal cost paths?  What i'm looking
> ] > for is the ability to use lsp load balancing without
> ] > having to static an lsp for all possible paths.
> ]
> ] LDP will do this automatically, as it follows the IGP; there are no
> ] provisions in TE to do this, but it could certainly be done by an
> ] operator.
> 
> 
> for ldp, yes if it's at the first hop of the lsp tunnel;
> but i think he is asking for ECMP even in the middle of
> the ldp tunnels. thus ldp needs to assign multi labels
> to its upstream for a single prefix which has ECMP.

not true with LDP in a frame-mode environment; ECMP is inherent in the
architecture.  with LDP cell-mode or with TE, yes, you're correct.


eric