802.11A (1232 AP) Real World Transfer Speed & Power Setting Question

Unanswered Question

<p>Hi, I am configuring bldg-bldg bridge access over 802.11A (1232AP). Using 17dBi external antennas. The configuration is working well so far, but I am concerned about optimizing transfer speeds (primarily a windows TCP/IP network) and setting the power correctly (not too low, not too high). The AP's are essentially across the street from one another (50m apart), with line of site. The antennas are externally mounted patch antennas aimed at each other. I currently have the root-bridge (! non-root) configured as follows:</p>

<p>speed basic-6.0 basic-9.0 basic-12.0 basic-18.0 basic-24.0 basic-36.0 basic-48.0 basic-54.0<br />

 no power client local<br />

 power client 11<br />

 power local 11<br />

 channel dfs<br />

 station-role root bridge  ! OTHER AP = station-role non-root bridge<br />

 antenna receive right<br />

 antenna transmit right<br />

 antenna gain 17<br />

 no cdp enable<br />

 bridge-group 1<br />

 bridge-group 1 port-protected<br />

 bridge-group 1 spanning-disabled</p>

<p>1) 300 MB file transfer takes about 5-6 minutes over the network. This is compared to about 1 minute over a standard 100Mbs network. No other hosts were using Wifi during this test. 2) The radio statistics show primarily 54 Mbps communication (95%). 3) The radio statistics (viewed through the Web GUI status) are attached.</p>

<p>Here are the questions:</p>

<p>1) How do I set the power optimally for this environment? I think I do not want to set it too high, but do not know how to measure "what is too high". How do I do that? Does it matter if the power is set to max with this small-range deployment?</p>

<p>2) My empirical file transfer test showed the Wifi link transfers files about 5 x slower than a 100 Mbps Copper network. Well, since 802.11A = 54Mbps and Cat5 Copper = 100 Mbps (in this test), you can probably see my question coming: 5x versus 2x? Is the performance loss due to the Wifi overhead? Is this loss reasonable?</p>

<p>3) Are my error stats within reason? Any built-in (1232AP) tests that I can do to measure the Wifi link quality? There is a carrier busy test, but I do not know exactly what that gets me.</p>

<p>thx in advance, W</p>

<p><br />

1) wifi_stats.doc</p>

Attachment: 
I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
scottmac Sun, 02/01/2009 - 06:06

The transfer speeds sound about right. The "54Mbps" is a signaling rate, not a throughput.

To make 802.11 wireless "reliable"  (comparable to a wired network)  the data is, in effect, sent twice and staggered such that a glitch usually doesn't get both.

In terms of throughput of your data, a strong signal with good signal quality, using IP, unencrypted  should run ~22-26Mbps (some variability for noise/interference, mixed frame sizes, TCP ACK times, application responses, etc). 

So, at ~24 Mbps (megabits per second) you're looking at ~4  megabytes per second versus 100Mbps/12.5mBps as a probable max rate.

Given that, a transfer that takes approximately one minute on a wired network under typical conditions ... having it take four-to-five minutes on a typical wireless system is about right.

 

For power settings, you can adjust the power by monitoring the RSSI values on the receiving system. If I can find the docs on Cisco's main site I'll post 'em up later (gotta run ...), but if the mechanical install is good, then it'll just be a little keyboard work.

Good Luck

Scott

 

Actions

This Discussion