RRM not able to bias power on cisco 1242 AP

[francis Letteri]

I recently installed a network of 750 Cisco 1242 LWAP's. A post installation RF survey revieled that AP:confused: 's with in close proximity were screaming at 100mw. The RRM feature was enabled on the controller. If you turned them down manually they would bias right back up. Anyone experiencing the same issue. The controllers are Cisco 4404's and WISM blades.

[golfnut]

That could be if it's only two or three APs wthin range (3 channel, non-overlap).

I'm not exactly familar with the Cisco AP switch however, you might want to check the CLI list. Other AP switches like Aruba have commands that will limit the -range- at which the APs will perform dynamic power adjustments.

Greg

[spiderbite]

Greetings,

Unfortunately they do that. What happens is that when the AP’s start up for the first time, they are all at 100 mW. What is supposed to happen is that as they hear each other they will all start to power down to their respective, correct power settings and dynamic channel assignments.

The problem is that they don’t.

The Transmit Power Control algorithm is responsible for powering down AP’s. Coverage Hole Detection is responsible for powering them up. It’s very fluky at best in my opinion. TCP usually doesn't work as advertised and CHD works too well and causes more problems.

The original Cisco document I saved says that the controller attempts to regulate everything so the third loudest neighbor is at -65dBm.

This would suggest that you have 50% overlap on your AP’s. The edge of your first cell (-65dBm) is supposed to be the location of your third AP (third loudest neighbor).

My deployments aren’t that dense usually. I’m good for 20 to 30 % from one AP to the next. Which is why the controller keeps the power maxed out trying to hit that mark.

The new Cisco document on the website now suggests the fourth loudest neighbor be at
-65dBm. This in my opinion is insane. That would be like 80% overlap! Cisco apparently got tired of people calling them and asking why Auto RF doesn’t work.

From my original document I saved:

Transmit Power Control Algorithm
The TPC algorithm, run at an un−modifiable ten−minute interval, is used by the RF Group Leader to determine APs_ RF proximities and adjust each band’s transmit power level downward, as necessary. This is performed in order to limit excessive cell overlap and co−channel interference.
Note: The TPC algorithm is only responsible for the downward adjustment of power levels. Increased transmission power is a part of the Coverage Hole Detection and Correction algorithm function, also explained in this document.
Each AP reports an RSSI−ordered list of all neighboring APs. Provided an AP has three or more neighboring APs, the RF Group Leader applies the TPC algorithm on a per−band, per−AP basis in order to adjust AP power transmit levels downward such that the third loudest neighbor AP is heard at a signal level of −65dBm (default value) or lower.
Power changes are only made when the third loudest neighbor of a given AP is heard at a signal level higher than the default value of −65 dBm.
Note: When all APs boot up for the first time, the APs transmit at their maximum power levels. Also, if the APs do not see each other with a −65dBm (default value), they push the power level up to the maximum in order to achieve this. If you go to the Controller CLI and issue the show ap auto−rf 802.11b<AP NAME> command, there is a Nearby APs section near the bottom of the command output. In this section, you find the dBm listing. If this dBm listing has values more than −65, then the AP transmit power is at the maximum. This is tweakable to an extent. You can issue the Config advanced 802.11b tx−power−control−thresh <−50 to −80> command. This command is discussed later in this document.
Note: When APs sustain a power cycle, the APs use their previous power setting. TPC adjustments
Subsequently occur as needed. See Table 5 for supported AP transmit power levels.


This is from the new document on the website:

http://www.cisco.com/univercd/cc/td/doc/product/wireless/control/c44/ccfig40/c40rrm.htm#wp1104178


Dynamic Transmit Power Control
The controller dynamically controls access point transmit power based on real-time wireless LAN conditions. Normally, power can be kept low to gain extra capacity and reduce interference. The controller attempts to balance access points such that they see their fourth strongest neighbor at an optimal -65 dbm or better.
The transmit power control algorithm only reduces an access point's power. However, the coverage hole algorithm, explained below, can increase access point power, thereby filling a coverage hole. For example, if a failed access point is detected, the coverage hole algorithm can automatically increase power on surrounding access points to fill the gap created by the loss in coverage.

Notice there is no longer any mention of the CLI command to change the TX levels. I tried this by changing from the default of -65dBm to -75dBm, but it didn't help much. (Full disclosure - the environment was surveyed for coverage not throughput, so I am still fixing that.)

Cisco wasn’t of much help at the time during the TAC case.

What I have done so far is manually configure my power and channel settings and turned Auto RF off. It fixed my issue but kind of negates the whole point of buying all this crap.

Needless to say, I am not a fan. However, I am currently resurveying vast amounts of real estate and most of it has not been deployed yet. Maybe once it is all in, I may give it a shot again, but the thought of going back through and killing Auto RF and manually set several hundred AP’s all over again does not excite me.

Please go check out the document and make up your own mind. Also TAC is your friend, so open a case and be prepared to be underwhelmed. Check out the Cisco Net Pro forums as there is much info there on this subject.

[spiderbite]

Other things that it wont do include but are not limited to:

Allow you to change data rates seperately. Since we are still fixing everything, there is a large quantity of AP's currently allowing 1Mbps traffic. The areas I have fixed were designed for 11Mbps at -65dBm. However, I can't certify the network since I can't change these datarates without changing the others. If I change the others I am sure to have coverage holes.

Show an accurate representation of your heatmaps. This is all show in my opinion. The coverage map has no idea what the walls are made out of and no way to tweak your findings once you have surveyed the area. That big green blob means very little in actuality.