[rbridge] Consensus Check: Announcing Root

[Radia.Perlman@sun.com]

Good point. There might be no bridges. So there has to be a way of encoding that. Anything such as:
a) leaving out the TLV in which one announces the root bridge
b) using a flag to say "no root"
c) using a special address, say 0
or I'm sure there are lots of possibilities.

Radia

----- Original Message -----
From: Eastlake III Donald-LDE008 <Donald.Eastlake at motorola.com>
Date: Saturday, October 6, 2007 12:46 pm
Subject: [rbridge] Consensus Check: Announcing Root

> This is a check via the mailing list on a slight modification of an
> apparent consensus from the minutes of the Chicago meeting for a 
> changefrom protocol draft -05. The tentative consensus at the 
> Chicago meeting
> was:
> 
>   It is mandatory for an RBridge to announce the bridge root that
>   it sees out each physical port.
> 
> Based on mailing list discussion, I would like to tweak this as 
> follows:
>   An Rbridge MUST parse BPDUs it receives on a port and announce
>   in the core IS-IS instance the bridge root that it sees out
>   each port. For MSTP (Multiple Spanning Tree Protocol), this is
>   the CIST (Common and Internal Spanning Tree) root.
> 
> I have a question in connection with this. What is announced for the
> port if no BPDU has been received recently or ever? Should there be a
> "root MAC address valid" flag per port or should there some 
> conventionalvalue, like zero, which is announced?
> 
> Thanks,
> Donald
> 
> _______________________________________________
> rbridge mailing list
> rbridge at postel.org
> http://mailman.postel.org/mailman/listinfo/rbridge
>