I know canopy says we can register 200 SM to an AP, but in reality in your experiences what is the limit before you start seeing degraded performance ?
We are hitting 100-130 on a few, latency is high, bandwidth is low, I think that is too many SM for it to handle efficiently…
whats your experience ?
If you want to keep “DSL-like” service the realistic load for an AP is around 80. After that you will start seeing performance drop off.
Aaron
does adding VLAN on an AP add more load then a standard SM…
so is that 80 SM regardless of if they are NAT enabled/disabled or VLAN enabled/disabled ?
when the SM is nat enabled then the RF will need to associate only one entry of MAC/IP to that SM, when nat disabled then the SM acts as an interface for may IP, in that case the RF needs to now learn about the internal network of the customer, hence adding more load on the bridge tables…
just guessing !!!
I don’t think it has to do with the bridging tables, but more so with the traffic that the processors can handle - both control traffic and user traffic.
Canopy Tech Support would be better suited to answer at this point.
Aaron
Tech Support … lol…
It’s a good place to start. Unless someone else on here can elaborate more than I can.
Aaron
If you have not called tech support, I highly recommend it.
I have been stumped a few times, and every time I have called my question has been answered promptly and accurately.
called and they went in circles… to what the documentation says and it all depends on what those 200 registered are doing…
its like they were being very careful not to say something that may contradict the manual…
all i want to know is what do you recommend as the upper limit… i don’t have any intentions of suing anyone…
One way to find out what is going on would be to deploy CactiEZ and start monitoring the bandwidth and ping latency on all the AP’s, BH’s and Routers. If it get beyond what you are comfortable then it’s maxed.
Another guage is when people start to complain…
i’m definately maxing them… just trying to gauge how much I should load them…
latency is the problem… not bandwidth… although due to bandwidth being calculated on an optimum size which is hardly achieved we are most likely running out of pps throughput…
switching to Advantage platform, hope that makes a difference.
We are not loading nearly as much as you are, but the Advantage definately helped.
awaiting custom clearance… so will hopefully deploy over the weekend…
two clusters of 4 AP’s