SXT 5HnD asymmetric throughput

Hi,

I have a problem with a wireless link using two SXT 5HnD connecting two (all switches are GigE Switches, all LAN Ports except the SXT 5HnD ones are 1Gbps) LANs using subnets 192.168.1.0/24 and 192.168.173.0/24 respectively. The distance is very short (on Google it shows 168m). The link is stable and when I use the built in bandwidth test I get full speed in both directions: When selecting 20MHz channels I get over 100Mbps in either direction and when selecting 20/40MHz HT Below I get over 200Mbps using UDP test with UDP Packet size 1500 and 500bytes.

However when I try to send data between 2 stations on the LAN I get asymmetric results. In one direction (remote → local) I get full 100Mbps throughput. In the other direction I only get approx. 35Mbps. (Using either FTP, iPerf or Ixia Chariot does show the same result)

I checked various stations on both LANs and regardless of the combination I always get the same results. The LAN stations are definitively not the bottle necks (in the most preforming setup they are a Quad XEON and Quad i7 running MacOSX 10.6).

I also tried connecting the stations directly to the SXT 5HnD. No change.

I also swapped the Mode of the SXT 5HnD (bridge / station bridge). No change.

Has anybody an idea what could cause this asymmetric behavior or any hint where I should look at?

Regards

Peter

What is the CPU usage like on both SXT’s when performing these tests?

reaim the antennas, even if you get a worse signal. you might be getting reflections / bounces on one unit. try moving one of the units a few feet away from where it is now.

if you are only 200m away from the other side you probably need to turn down the power.

When performing the good direction the highest CPU I have observed on the station side is 77%/73% and on the bridge side 60%/59%. In the bad direction with only 35Mbps the highest CPU I have observed is 37%/32% and 18%/17% (CPU/IRQ). Lower throughput seems to have another reason. But thanks for the hint, I did not even check that before.

The point with the antenna/reflection and the power is very good. I’ll try another position/angle and settings for power. Thanks as well. I’ll keep you informed about the results.