lower than expected 1310 bridge performance

[igor.sotelo]

Hi All,

We have recently installed an wireless WAN link over 7 km (aproximate 4.5 milles) distance. The towers are sufficent for fresnel and earth bulge and obstacles requirements.

We have used Cisco Aironet 1310 bridges (one configured as root bridge, the other as non-root bridge). The antennas are 24 dBi from Hyperlink, connected by RP-TNC to N pigtails, also provided by Hyperlink. According the Cisco utilities, the antennas has more than enough gain for the distance, climatic and topographic conditions.

The alignement of the antennas was made visually (using binoculars) on one side, on the other with precision instruments.

To select the frequency (the automatic selection works awfuly bad, it establish connection one of three times) we use the Carrier Busy Test on the network interface / Radio 802.11G menu, where we are able to pick the frequency with zero percent (0%) utilization most of the time.

The problems we experiment are the following:

- The latency times are very variable. From 1 ms, to 207 ms, to other values. We have used other wireless equipment with more stable performance.

- There are lost packets frequently, even with the slightest network traffic (2 PC's with terminal service sessions). Supousedly, the link is of 54 Mbps speed, but the quality is very low.

- At the moment, we are losing connection every couple of minutes. After that the non-root configured bridge stays down and the only way to reestablish connection is by reseting the 1310.

- I see that the speed changes every time I refresh the IE utility.

I still havent configured any security, because the awful way the link works.

What could be the reason for such bad performance?

Best Regards,

Igor Sotelo.

[kbschindler]

The first thing that I would do would be to take bridges off of the network. Connect a laptop directly to the bridges on either side configured with IP addresses and do a constant ping. This will tell you whether it is a network issue or an issue with the bridges. If the bridges stay up and the pings are good, you need to look at the network side of it.

I had a customer a couple of years ago with a similar problem using 1400 bridges. They ended up having a misconfiguration on a switch. Unfortunately, I don't know what it was, they never got back with me.

I hope this gives you a good start.

[igor.sotelo]

Hi All,

I have noticed that when the switches have VLAN configured, the latency is very variable. In this particular case, the switches does not have this configuration. I haven't configured VLAN's on the wireless bridges either.

I will test the link without any switches, it's an good idea. Perhaps there is "something" with the wired network.

Other than that the specifics are:

- I have used Belden RG-6 1530A cable that has even higher grade than the Belden 9077 recommended by the manuals.

- The extension of the cables is around 40 meters (120 foot).

- We have installed another wireless link in the place, that uses the same frequency, but different polarity. Also we try to separate the channels at least 7 frequencies.

- The other system is omnidireccional in nature, and it doesn't have excessive gain.

- On the place where the equipment of both systems coexist, the physical separation between the antennas of the two systems is around 9 foots (3 meters).

The error messages we get are:

- On the root bridge:

Mar 1 00:00:57.238 Information Interface Dot11Radio0, Deauthenticating Station 0017.0ec6.a590 Reason: Previous authentication no longer valid

Mar 1 00:00:57.237 Warning Packet to client 0017.0ec6.a590 reached max retries, removing the client

- On the non-root bridge:

Mar 1 16:19:52.072 Notification Line protocol on Interface Dot11Radio0, changed state to up

Mar 1 16:19:51.072 Error Interface Dot11Radio0, changed state to up

Mar 1 16:19:51.071 Warning Interface Dot11Radio0, Associated To AP Central-2 0017.0ec6.a580 [None]

Mar 1 16:16:00.255 Warning Interface Dot11Radio0, cannot associate: No Response

Mar 1 16:15:50.397 Notification Line protocol on Interface Dot11Radio0, changed state to down

Mar 1 16:15:49.398 Error Interface Dot11Radio0, changed state to down

Mar 1 16:15:49.397 Warning Interface Dot11Radio0, parent lost: Too many retries

Mar 1 16:14:56.788 Notification Line protocol on Interface Dot11Radio0, changed state to up

Mar 1 16:14:55.788 Error Interface Dot11Radio0, changed state to up

Mar 1 16:14:55.788 Warning Interface Dot11Radio0, Associated To AP Central-2 0017.0ec6.a580 [None]

When we make a reload on any of the bridges the link reestablishes for some time. With this message on the root bridge:

Mar 1 00:00:44.194 Information Interface Dot11Radio0, Station NONROOTNAME 0017.0ec6.a590 Reassociated KEY_MGMT[NONE]

Mar 1 00:00:35.456 Notification Line protocol on Interface Dot11Radio0, changed state to up

I will appreciate any additional help.

Best Regards,

Igor Sotelo.