[rbridge] (no subject)
riw at cisco.com
Thu Dec 6 19:06:29 PST 2007
-----BEGIN PGP SIGNED MESSAGE-----
> The point is most bridges don't have scaling issues like
> that. If we're going to build a technology that costs
> as much as routers and has the same (if not even worse
> scaling properties), why would people want to deploy it?
I would argue that most bridges do have scaling issues like that, first
of all. Second, I would argue that rbridges, built properly, and
deployed properly, won't have worse scaling properties. Finally, I'd
point out that rbridges might not replace STP bridges, but rather
augment them in some specific places, etc.
In general, I consider this line of reasoning a red herring. We need to
focus on realistic situations, and find solutions which don't incur
large new pieces of complexity for very small gains. I would also point
out that if you're counting on this "elect a single DRB on a segment"
idea to prevent problems in the case you're using as an example--it
won't in any case. I've already shown two cases where it fails, and can
easily find at least one more. Those cases apply to your examples, as well.
riw at cisco.com CCIE <>< Grace Alone
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----
More information about the rbridge