[c-nsp] purposefully mismatching native vlans

[Andrew Fort]

sthaug at nethelp.no wrote:
>> I'm not asking rhetorically here, i'm hoping somebody can tell me why
>> disabling spanning tree is so compelling.  I've seen a carrier with a
>> bunch of Ethernet out toward their customers disable spanning tree, and
>> I just dismissed it as goofy, but now with this thread I'm starting to
>> wonder if its me that has it all wrong.
> 
> Here are a couple of reasons why you might not want to use spanning tree
> towards your customers:
> 
> - It doesn't scale. Or at least per-VLAN spanning tree doesn't scale.
> - You really don't want to mix customer STP traffic with your own STP
> traffic, this is a recipe for disaster.
> 
> Steinar Haug, Nethelp consulting, sthaug at nethelp.no

Hear, hear.  To add to this: Can Cisco bring on provider bridge and 
_useful_ (read: minimum 2 per port) VLAN translation, and make us native 
Ethernet providers happy!  We'd even consider buying more of your boxes! 
   Before someone says "VPLS", is it really the answer for an ethernet 
service provider, or is it just a multipoint core technology?  (scaling 
limits on every VPLS vendor's kit sofar would tend to make it useful 
only for the latter).

-andrew