Ok, so after I figured out what I described in my previous thread, my next assignment was to measure speed on 802.11n networks.
To do this, I used a 802.11n wireless USB card with RT2870 chipset and/or two RB800 RouterBOARDs with their bundle antennas. The routerboard was connected to a 100mbit ethernet with a fixed IP, and wireless clients connected to it through NAT. Internet worked on all wireless clients.
Try one:
I configured the RB800’s wireless card (wlan1) in ap bridge mode, 2GHz-B/G/N wireless mode, on channel 2, using both antennas for send and receive, HT extension channel ‘below control’ (I wanted 40MHz channel width) and AMPDU priorities all checked. The goal was to achieve highest possible thoughput.
I connected from a laptop to the AP. The declared speed in Ralink configuration utility was 300Mbps and internet was available. Prior to this I put a 512mb file to a local computer in the network. When I tried to fetch a file with wget, the speed was at best 2.3mb/s but it was jumping all around, even dropped at 100kb/s few times.
I concluded network was choked and that speed drop was due to unavailable ethernet bandwidth (it’s a very large college network).
Try two:
This time I turned off the laptop and turned on a second RB800. I configured it in station mode with a different IP than the first routerboard, found the first AP with Scanner and connected to it. It appeared in Registration as connected client.
Then I turned on Bandwidth test on a second one, entered the ip of the first one and chose udp, entered credentials of the first one and started the test. The result was it showed at best 3Mbits in both directions, while average was around 1Mbits?! I really don’t understand how N networks could be this slow, something is obviously wrong. The speeds reported when I look at statistics of the client are known to jump to 54Mbit, very rarely a bit more, but they usually stay fixed at 11Mbit or less.
So next thing what came to mind is that I misconfigured something. So I played with all the settings from the above, I tried to put both APs in N-only mode, tried seperate antennas, disabled HT extensions, etc. etc. I can’t even remember all the settings I tried.
So, what’s the deal here? My assignment was to determine how 802.11n networks are faster than 802.11g networks.
The conditions were: the routerboards were a 1 meter apart, in an environment with G networks available, but on different channels. Channel 2 which was used by both routerboards was free.
Thanks for any help!
P.S. I won’t have access to the routerboards for a week or so, which gives me time to figure this out but won’t allow me to test any of the suggestions 