Hi all,
I'm trying to get my head around the queuing changes in ROSv6.
I've spent some time thinking about what I want to do, and I think I've got a plan. I'd like to run an approach by someone more knowledgable than I to get a second opinion.
I'd like to implement two classes of service (business and residential) with different contention ratios. Residential at 1/15, business at 1/10. Business and residential CPEs are mixed with IPs in the same /16 subnet.
I understand that with the efficiency changes in simple queues, and the goal of doing double qos later that simple queues is where I should do my per-subscriber queuing.
My plan is this:
-create a top level parent queue with target=0.0.0.0/0 and dst=<WAN interface>, max-limit up/down =total bandwidth available from upstream
--create a child queue with target=10.40.0.0/16 (customer subnet) and dst=<WAN interface>, max-limit up/down =total bandwidth allocated to customers
---create child queues for each business class customer. target=<customer IP>, dst=<WAN interface>, max-limit up/down=<subscribed bandwidth> limit-at=<1/10 subscribed bandwidth> priority=2
---create child queues for each residential class customer. target=<customer IP>, dst=<WAN interface>, max-limit up/down=<subscribed bandwidth> limit-at=<1/15 subscribed bandwidth> priority=3
In my understanding, this should:
-guarantee business customers 1/10 of their max bandwidth and residential customers 1/15 of their max bandwidth
-when extra bandwidth is available, it would be given to taking business class customers (priority 2) to their full max-limit and any left over would be allocated to residential customers (priority 3)
Does this sound right? Am I missing anything?