1410 bridges - poor throughput

[ED CARMODY]

Hi, I have two 1410 bridges installed 3 kilometers apart, using integrated antennae. I have measured the throughput across the link using two PCs running Qcheck that are directely connected to the bridges, and am only seeing ~15 mbps towards the root bridge and anywhere from 4-12 mbps away from the root. I've flipped the root/non-root configs on the bridges, and the slowness went in the other direction

Both bridges have consistent RSSI at -54 or better. Both bridges' carrier busy tests show pretty clean, with only one location showing utilization on 5745 @ 13%, so I set the bridges to use 5785.

I do see that the link seems to be stepping down from the 54 mbps data rate, down to 48, 36, and 24, with what I percieve to be high data and RTS retries on one of the radios, which makes me suspect interference, but carrier busy is clean.

Experimenting, I have tried hard-setting the data rates on the radios, and have found that the radios cannot maintain the link when I hard-set the speed to anything above basic-24.0 (i.e. at basic-36.0, the link bombs).

Any ideas on why throughput might be so low, and why the link is faster towards the root than away from the root?

What should concatenation be set to? What should CCA be set to? RTS threshold? My customer is trying to move 10MB graphics files over the link in rapid succession (PACS).

Thanks in advance.

[steve20]

Just an FYI. I am having the same exact problem and have tried the same hard-settings with the same results. Will let you know if I have any luck on my end!!!

[steve20]

Do you have the FastEthernet and switches set to 100MB/Full? If so, change to auto on both. Seemed to fix my problem. Also try different frequencies.

[ED CARMODY]

Switch ports are set to auto already, and have tried various frequencies, with no difference in performance.

We have run carrier busy tests to select a clear channel, but I'm still seeing significant "energy detect defers"; doesnt' this indicate interference? Why wouldn't it show up in the carrier busy test?

[abrahamj]

Hey E,

I have a dozen links supported by 1410's. They vary in length from 3 to 8 miles. I do not use integrated antennas, in favor of parabolics, because I believe that the coming congestion at 5.8 will kill anything but the most tightly focused xmit signal.

Here is a link to a cisco doc on 1410 configuration: http://www.cisco.com/en/US/customer/products/hw/wireless/ps5279/prod_release_note09186a00802146c0.html

It says that the default settings for CCA are 0 and 62, although I have never had anything go to 62. TAC has told me that 61 should be the default on all bridges (1310,1410) so I have it coded. I haven't messed with concat or RTS threshholds.

I have experienced your problem and found that going to a different pole (vertical or horiz) to be the best solution. Since getting real hands-on support is such a challenge here, feel free to email me at abrahamj@saccounty.net if you ever need any help. I have about 50 bridged links in and around Sacramento, CA - using Cisco, WMUX, Proxim and Orthogon - and will be happy to offer help where I can.

[ipbod1]

Hiya,

I am running/testing 2x 1310G in a bridge set up here. Putting the 2 antennae next to each other I managed to get approx 22/24 mbs at max. I then decided to place them in position. 1 in the main building and the 2 in the opposite building ( 30 metres away) and the rate dropped to 3mb at max!!! which really is a big difference. We are using 8db antenae and just brought 12db ones which were worse :(

Any ideas how to improve this or the cause?? The Windows maybe in between buildings??

2. What is the maximum data rate I should be getting from these say if i put them next to each other. Am using iperf to test this.

We had a wireless survey done by specialists and was found the site to be good for wireless connectivity due to the close proximity and DLS.

Any ideas would be appreciated??

IP-bod