Network Busy on 7920

danjor · ‎07-07-2005

Cann any one tell me what exactly mean NW busy on the phone display when geting the line after enering the dialing number?

It seams that we do not have to mutch IP phones on the AP where the phone is assoc butin one region from NW we do have those error very often.

michaelgillespie · ‎07-08-2005

If you get "Network Busy" when dialing out, then this means the CU is >= 45. CU (channel utilization) is derived from QBSS that AP sends out in it's beacons. If this is 45, then there is significant across that AP. There was a bug w/ 12.3(2)JA though where this was falsely advertised. Should upgrade to 12.3(4)JA. We use QBSS for outbound CAC.

Tobias Heisele · ‎10-26-2005

I updated from 12.3(4)JA to 12.3(7)JA1 and have the same problem now. Is there a bug in the new release, too?

danjor · ‎10-26-2005

We did an update to the 12.3(4)JA from an older version and the error did stop (but in the same time we did also add one more AP in the region we had the trouble). After this update the errors stop. No test with the new release 12.3(4)JA1 and not sure if the update did help or if this new AP did the job.

michaelgillespie · ‎10-26-2005

More than likely you are load-balancing the traffic now across the 2 APs. Only 1 fix in 12.3(4)JA1.

Also in 12.3(7)JA, there is a new QBSS, which is CCA based. This on 255 scale. Will get "Network Busy" on outbound calls if the QBSS value is >= 105. Should use 7920 version 2.0 or higher.

Tobias Heisele · ‎11-18-2005

We are using 7920 firmware 2.0, but on some special places, we got the "Network busy" message. When I downgrade the local ap from 12.3.7-JA1 to 12.3.4-JA1, the phones work fine.

michaelgillespie · ‎11-19-2005

So QBSS in 12.3(7)JA or higher is CCA based. Will show value presenting "real" channel utilization not just a value for how much traffic is currently on that AP's radio. If is high (100+) w/o any traffic, then may need to use another channel.

Tobias Heisele · ‎11-28-2005

I reconfigured the channels twice using wlse assisted connfiguration feature, but the problem changed only its location. Is there a way to ensure the wirless phones become working?

michaelgillespie · ‎11-28-2005

You need to look at your spectrum. Ensure you are using 2.01. 12.3(7)JA2 is latest AP code on CCO. Can't assume all 802.11 channels are usable / available. If you get high utilization #s (i.e. >= 100) w/o any 802.11 traffic then you may need to take that channel out of the design. Spectrum analyzer can help track down non-802.11 energy that could be potentially interfering. Airmagnet now OEMs a very inexpensive version from Cognio. Temporary workaround could also be to turn off QBSS, but then phones could get overloaded on an AP, voice quality could get impacted, etc.

danjor · ‎10-26-2005

As we have only 6 fix IP Phone and 40 Phone 7920 the impact from those trouble was very high in our NW. But we never get a final answer. Fror Cisco or other CCIE it was folowing as most often answer:

- when the CU on the phone Site survey is biger than 45 (meaning the CPU usage)

- Case extract: must have an RSSI that is greater than 35 (which is equivalent to a receiver threshold of -67 decibels per milliwatt) and a channel utilization QoS Basis Service Set (QBSS) load that is less than 45

- To many bit errors and bad quality from WLAN is given those error

In fact we had for example a Aironet1100 where one phone was assiciate. This phone was not on call but the CU was still on 200. As this parameter is explained as a % CPU it is not possible but any way, it the explanations I get.

Some user stays under an AP with 2 or 3 associations and get also the newtwork busy.

The probleme was located on 2 AP from 5 installed.

After the upgrade to 12.3(4)JA , no more Network busy.

michaelgillespie · ‎10-26-2005

So if the CU is 200, then this is CCA based QBSS and this channel is basically in a DoS state. Even a microwave will only register to about 130 for QBSS (like 10 calls on the AP).

There was an issue before in 12.3(2)JA, where QBSS was CCA based on 255 scale when it shouldn't have been. Would not be able to make many calls at all w/ that image. There was a DDTS for this issue. It was fixed in 12.3(2)JA2.