Hi,
I am a service provider and have customers largely terminated via PPPoE
on a 7201. I typically enforce maximum upstream and downstream rates at
the customer's CPE and charge for higher rates. This works reasonably
well, if inflexible. I'd like to move to a more intelligent model which
would allow me to implement some sort of dynamic QoS for my customers
where I can demote users who are over their limit to a lower QoS tier
and/or lower bandwidth profile.
I have radius of course and already can insert various attributes to
the customer and have some already setup with their own interface
bandwidth limits and such. The problem is if I base my model on putting
out radius attributes, I have to force the customer to disconnect and
let PPPoE re-negotiate each time the customer is promoted or demoted.
This seems unfairly disruptive to me.
I would also like feedback / opinions on the scalability of
implementing rate controls on the virtual-access interface for each
customer. Is rate limiting really going to work per-session with 4000+
pppoe sessions on the same router?
Mike-
_______________________________________________
cisco-nsp mailing list [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
