[rbridge] loops in trill networks
Larry Kreeger (kreeger)
kreeger at cisco.com
Tue Oct 11 11:52:25 PDT 2005
This brings up another question.
What does the RBridge network look like to the 802.1 bridged network.
Does the RBridge network pretend it is a giant 802.1D bridge -
exchanging BPDUs at its edge ports with the 802.1 bridges and blocking
edge ports if it detects a loop? What happens to the BPDUs sent by
802.1 bridges to the RBridges?
From: rbridge-bounces at postel.org [mailto:rbridge-bounces at postel.org] On
Behalf Of Joe Touch
Sent: Tuesday, October 11, 2005 8:26 AM
To: Developing a hybrid router/bridge.
Subject: Re: [rbridge] loops in trill networks
-----BEGIN PGP SIGNED MESSAGE-----
Loa Andersson wrote:
> Oh, yes it is true ...
> but I thought that a trill network could consist of a mixed of
> rbridges and the type of bridges that exist today, if that is the case
> you can possibly form the loop over only non-rbridges
The conventional bridges have a spanning tree, which is what already
prevents such loops currently.
Think of the path between two nodes (hosts, routers, etc. - anything
that sources/sinks L2 frames) as having three components (at most):
the ingress (bridgepath1) and egress (bridgepath2) conventional bridge
paths use spanning trees, so they can't have loops.
The rbridge uses the TTL, so it can't have a loop at its layer. Even
when it tunnels over conventional bridges ('inside' the rbridge, in a
sense), those can't have loops because they're spanning-tree.
The combination of spanning tree in all bridge paths and TTL in the
rbridge logical path prevents loops in the node-node path.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (MingW32)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----
rbridge mailing list
rbridge at postel.org
More information about the rbridge