[rbridge] Orphaned endnodes with partitioned VLANs on a cloud
anoop at brocade.com
Thu Dec 6 19:46:30 PST 2007
> -----Original Message-----
> From: Russ White [mailto:riw at cisco.com]
> Sent: Thursday, December 06, 2007 10:42 PM
> To: Anoop Ghanwani
> Cc: Eric Gray; Radia Perlman; Developing a hybrid router/bridge.
> Subject: Re: [rbridge] Orphaned endnodes with partitioned
> VLANs on a cloud
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> >> We should *never* assume an underlying media of any type.
> It will tie
> >> the spec into a corner, and make it impossible to use in other
> >> situations.
> > This I do not get at all. I need more explanation with an example
> > that does not work.
> I would think it should be clear.
> Rbridges should care about getting good connectivity between
> the rbridges, not fixing bad connectivity for this media type
> or the other because they might happen to be the media type
> used to connect the rbridges. Should we also be specifying
> the way an rbridge interoperates with an ADM if there happens
> to be one in the middle? Or perhaps a satellite link?
> The connectivity between the rbridges should be assumed, and
> we should worry about how to emulate a bridged network over
> that connectivity. All this stuff trying to figure out how to
> make it work over a specific media in a specific situation
> should be put in another doc, if it's discussed at all, and
> not in the base specification.
If the world were really that simple we'd be done by
now. Reality of it is we have to worry about what's
going on in the underlying media. For every interface
type that you can run a router on, we have a corresponding
RFC that says what specific things need to be done for
that interface type. When we get to other RBridge
interface types, we'll have to deal with similar issues.
More information about the rbridge