PMP 450M Contention Slots mixed with PMP450i Auto

What are you geeks doing with contention slots when you have a mix of 450M and 450i units running on the same tower? Using auto for the 450i units but you all know that isn’t an option on the 450M units. I just set it as 6 for now.

To take it further, I am not sure the document for contention slots is useful when it comes to 450M units in general. For my example, assume we run 40-50 subs per AP running 6 as the control slot. On 20Mhz I get a max of 140 Throughput and a maxed downlink frame on just about every AP during peak. All subs are 8/6 or above. Seems like this should be a better performer.

Even when using auto on 450/450i, you’ll still want to make sure your contention slots match on 450m. The value is still used for frame sync purposes.

From what I understand, guidelines for contention slots applies the same for 450/450i/450m.

1 Like

Yes, even when configured to use Auto Contention on 450/450i/MicroPoP the “Contention Slots” parameter is used for co-location purposes. So yes, if using the same Channel Bandwidth, Frame Period, Max Range then can just have Contention Slots match your 450m. If not however, use the Frame Calculator page and make sure one AP’s Antenna Transmit End doesn’t overlap the other AP’s Antenna Receive Start time and you’re synchronized.

Yes, this document doesn’t specifically mention 450m, but the same concept applies as with 450/450i/MicroPoP before Auto Contention setting. Basically the more Control Slots you configure, the lower your top line throughput can be (sacrificing slots to contention) but the more steady your uplink traffic will be. Of course with TCP ACKs you need uplink to be consistent to get downlink throughput. But if even configured to the minimum Control Slots and uplink traffic is maxed and then no Bandwidth Requests get through, the uplink becomes empty and then all the unused slots become additional temporary Contention Slots. In this matter it kind of works itself out momentarily. If you blasted traffic in a lab scenario with low Contention Slots configured, you see a bit of up and down uplink throughput graph. And the more Contention Slots the smoother and lower it gets.

450m makes this really hard to make any general statements for X SMs how many Contention Slots should be configured, as the 450m can receive multiple Bandwidth Requests in the same symbol at the same time if they are properly spatially separated. So it’s not just purely about Bandwidth Request Request Percentage but percentage received at your worst spatial frequency. Either way monitoring your Bandwidth Request Percentage will give you an idea how this is performing.

Are you talking about your 450m or 450i here? If 450m, what does your flood test results show? 140 Mbps?

1 Like

Any word on auto-contention for 450m hoped we’d see it creep in on 20.3

1 Like