The MPLS WG Archive

[Ajay Simha <asimha@cisco.com>]

On Sat, 29 Jun 2002, Shahram Davari wrote:

>SD: Silly or not, I am afraid it is the reality. LDP is inherently unable to
>SD:do load ballancing, that is why this kludge is there.

Shahram,

What are we comparing this against? IP? Please explain how in the IP world one
achieved IP vs non-IP load balancing and how it is different when you run LDP?

-ajay
>SD:
>SD:-Shahram
>SD:
>SD:-----Original Message-----
>SD:From: Eric Osborne
>SD:To: Shahram Davari
>SD:Cc: 'Eric Osborne'; George Sheng; scullptor@yahoo.com; mpls@UU.NET
>SD:Sent: 6/27/02 7:11 PM
>SD:Subject: Re: your mail
>SD:
>SD:On Thu, Jun 27, 2002 at 01:02:53PM -0700, Shahram Davari wrote:
>SD:> > ECMP is inherent in the architecture.
>SD:>
>SD:> What? ECMP is a kludge to overcome LDP's inability to TE.
>SD:
>SD:That seems to be an Extremely Silly statement; if you want TE, use
>SD:RSVP.  RFC3209, in case you've missed it...:)
>SD:
>SD:
>SD:
>SD:eric
>SD:
>SD:> If you think about it doing hashing of multiple labels and perhaps
>SD:> IP header in the middle of an LSP actually defeats the purpose of
>SD:> MPLS, which was supposed to be have a simple label lookup in the
>SD:> core.
>SD:
>SD:> -Shahram
>SD:>
>SD: