[quagga-dev 3765] Re: [Quagga-bugs] [Bug 220] New: ripd doesn't recognize ethernet peer to peer addresses under Linux

Alia Atlas akatlas at google.com
Tue Oct 25 18:59:13 BST 2005


Greg,

I'm not certain how it would apply for RIP, but I am familiar with the
following
draft: draft-ietf-isis-igp-p2p-over-lan-05.txt

This suggests that when Ethernet is deployed as point-to-point, then it
would
be desirable/more efficient to represent it as such. I was looking for a
similar
draft for OSPF, but don't see it off-hand.  I believe I recall hearing about
such.

Alia

On 25 Oct 2005 07:19:05 -0400, Greg Troxel <gdt at ir.bbn.com> wrote:
>
> I would suggest examination of the relevant standards.  While it's
> possible to configure interfaces in many ways, and to make quagga
> 'work' over those configs, for some value of work, my own preference
> is that a) quagga conform to standards and b) any extensions beyond
> standards be well thought out and documented.
>
> If someone wants to use Ethernet transport to create a point-to-point
> link, then I can use using addresses .1 and .2 in a /30, or perhaps .0
> and .1 in a /31 with no broadcast address, or defining a way to
> present this as a POINTOPOINT interface type.
>
> I'm all for cleaning up wrong assumptions, but currently there are
> well-agreed-on rules for the use of IP on Ethernet, and as far as I
> know those don't include having a 'peer address' configured.  Pointers
> to evidence of widely accepted (particularly interoperable) practice
> would be appreciated.
>
> --
>         Greg Troxel <gdt at ir.bbn.com>
> _______________________________________________
> Quagga-dev mailing list
> Quagga-dev at lists.quagga.net
> http://lists.quagga.net/mailman/listinfo/quagga-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quagga.net/pipermail/quagga-dev/attachments/20051025/22e65b61/attachment-0001.html>


More information about the Quagga-dev mailing list