Nanaimo <> Vancouver AirMux test
After weeks and weeks of preparation and conspiring the day finally arrived. Of course the night was fairly sleepless, there’s always some last minute bits that need to be worked out, like making sure your client access point bothers to boot up, or having a map printed 
We set out in the wee hours of the morning to the Nanaimo test site. Arriving in the early afternoon the three of us quickly setup the link equipment and local network. After a short debate on the merits of using the pre-calculated azimuth and inclination offsets vs. “The site’s over there, just point the damn thing” we opted for the professional method and eyeballed the antenna until it looked just right.
We left the mux in link installation mode while we continued setting up the rest of the local network, leaving it to beep once per second to remind us that it was indeed still operational and not catching fire in the hot afternoon sun. About 10 minutes or so later, we heard the tell tale double beep which could only be a sign that the Vancouver team was set up and ready to go.
Once the alignment was done we puttered around with the mast, tightening it ever so strongly to keep it from flapping in the wind. Within a few minutes we had achieved a –69/-70 signal level at 12 Mbps.
Our original test plans had called for setting up several services to demonstrate the prowess of the mux link:
- Live video conferencing between the two sides
- A loop back on one end, Cisco 3620 on the other to bridge two T1 channels
- Internet access
- Media streaming
Unfortunately we ran into some problems :/ . Our web cam in Nanaimo was fussy and had problems with the driver installation. After half an hour of screwing around with a brain dead driver disk various mutterings of an annoyed operator were heard and the sound of a “creativity-sustaining beverage” being opened echoed against the mighty dish.
Fortunately the Vancouver team did have their Sony camera up, we watched the view from Cypress over the link in near-real time.
We fed Internet access from a Linksys WRT54G running OpenWRT. The WRT had two 7 dB high gain antennas attached and was connected to our friends at the NETGEAR community network. Signal levels on the NCN link were approximately –85 dBm, but functioned at 5.5 Mbps with very little packet loss.
The AirMux link itself performed very well. We did not observe any appreciable packet loss or signal quality issues during the test, and we were able to establish an air speed of 36 Mbps with no degradation
We were, however, unable to maintain a 48 Mbps link across the link. Setting the air interface to 48 Mbps caused the muxen to lower their overall power output resulting in not enough receive strength to trip the muxens receiver threshold.
The TDM trunk test did not work so well, throughout the test the T1 WIC interfaces were in alarm and could not be coached into playing nicely. We suspect the problem may be a hardware incompatibility between the (newer) fast Ethernet module and the (older) WICs, although we did have the TDM layers working earlier in the week.
At 36 Mbps we measured a sustained IP throughput of ~5.7 Mbps, full duplex. This is pretty close to what the budget calculator had to say:
Rate
36 Mbps
Frequency
5.8 GHz / TDD / 20 MHz
Antenna gain
32 dB
Tx Power
14 dBm
Max Range
58 KM
Our Distance
65 KM
Expected RSS
-68 dBm
Ethernet Throughput (w/2xT1 TDM)
2.6 Mbps
Ethernet throughout (no T1 TDM)
5.9 Mbps
Well now, that’s pretty close to what we achieved, and at slightly longer distance than what the budget calculator rated the link for.
We also compared horizontal vs. vertical polarization. There was no noticeable variation in the signal.
All in all I was quite happy with the ease of setup and performance of the link. Alignment was pretty simple (granted we did have beautiful line of sight) and the throughput was pretty much what I expected, although not as speedy as say, an Orthogon, but the price is right and the performance is gravy.
