[rbridge] Time to summarize "forward or block" BPDU thread

[Joe Touch]

Peter Ashwood-Smith wrote:
>   This lack of understanding of the DRs seems to be the cause of a lot
> of this debate. 

IMO, it's the belief that there will only ever be one DR or DR-like
(BLOCKing) device. I agree that this all works when it works.

>   Also, the 'problems' that people are attributing to blocking and not
> processing BPDUs [BLOCK] are also 'problems' when no STP is running.

'problems'? (why the quotes?)

OK, let's just shut off spanning tree in bridges and it all works just
fine now?

Now I'm *really* confused.

AFAICT, this all works when there is at most one rbridge campus in an L2.

WHEN (note the use of 'when', not 'if', since there's *nothing* that can
be done to enforce this) this isn't the case, the following problems arise:

	a) there is no way to detect the error

	b) loops *will* occur

If we want to say "BLOCK *MAY* work, but *MAY* silently fail", and
"PARTICIPATE and TRANSPARENT will always work with current standards",
that's fine (and, AFAICT, true), but I also don't see how we pick BLOCK
as a default in that case.

Joe
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 254 bytes
Desc: OpenPGP digital signature
Url : http://www.postel.org/pipermail/rbridge/attachments/20051017/4d8444d5/signature.bin