802.11s mesh issue

General
1

Hi, I am currently using the HaLowLink 1 devices for testing the possibly of using Wi-Fi HaLow for communication between a base station and an autonomous ground vehicle to replace the UHF used today. The purpose is to allow for long range communication in areas without direct line of sight using the mesh capabilities.

To test this i used three HaLowLink 1 devices, one set up as an AP configured in the 802.11s mesh network using the wizard in the GUI, as described in the user guide. The other two devices were then configured as extenders (or mesh points) using the push button on the bottom of the devices, and paired to the AP. Before configuring the devices they were all updated to the most recent firmware (2.7.6).

Since EU is not supported yet, the channel was set to US with a width of 1MHz on channel 33, for the furthest range. Additionally, the 2.4GHz Wi-Fi is disabled for the two devices in extender mode, and the TX power is reduced to 13dBm on all three devices, by accessing all devices through SSH and using the command “iw dev wlan0 set txpower fixed 1300”. Note the TX power is changed after turning on each device, since this configuration is not stored on the device. The rest of the configurations were left untouched.

TEST 1: (AP ← ~600m → Ext1)

Once setup I found a location with line of sight and placed the AP and one of the extenders (Ext1) approximately 600m apart and ran the following bash script from the AP:

#!/bin/sh

#Check if a filename argument was provided
if [ -z "$1" ]; then
     echo “Usage: $0 <output_filename.json>”
     exit 1
fi

iperf3 -c 192.168.12.114 -4 -u -b 50k -t 60 -O 5 
     –udp-counters-64bit --dont-fragment --repeating-payload --get-server-output 
     –json > "$1"

where the IPV4 address matches Ext1. I ran the script twice and got a decent connection with the following stats:

  • Realization 1: Avg. bitrate 49.19kbps (packet loss: 2.70%)
  • Realization 2: Avg. bitrate 41.44kbps (packet loss: 16.22%)

TEST 2: (AP ← ~300m → Ext2 ← ~300m → Ext1)

Then without turning off or moving the AP and Ext1 i turned on the second extender/mesh point (Ext2), configured the txpower to 13dBm as described earlier, and placed it in between the other two devices approximately 300m apart from both in a straight line. Then i ran the bash script again from the AP and got the following results:

  • Realization 1: Avg. bitrate 32.21kbps (packet loss: 35.52%)
  • Realization 1: Avg. bitrate 35.69kbps (packet loss: 28.96%)

So somehow the bitrate is worse by incorporating a mesh point between the two devices? It is worth noting that the AP and Ext1 are both powered by identical power banks, while Ext2 was powered by my laptop, which I assume should not cause an issue.

I hope someone are able to help me understand why i see this results, and if it is just some simple configuration mistake that I made, or some other dumb mistake.

2

I’m not the expert on this, but a few thoughts:

  • by introducing the mesh point, you add another device on your channel competing for airtime, and some packets may be directed via it; adding a mesh when it’s not necessary is going to make things worse, particularly when you only have 1mhz bandwidth available
  • you use UDP and restrict the bitrate to 50kbps; I would have thought you’d get a higher bitrate than this, even with the restrictions you’ve mentioned. Did you select 50kbps because you noticed that this was the highest possible? Have you experimented with TCP?
  • if you don’t need the flexible topology of 11s Mesh, it may be worth trying EasyMesh, as it will likely deliver higher performance for a straightforward non-LOS use case.

You may also find Why does adding each new node result in ~50% bandwidth reduction, even when stations are in close proximity? - #2 by swadsworth interesting, though it obviously isn’t exactly the same situation.