Question on 802.11b channel assignments

I've got a remote site that has a total of 25 1131 access points under the control of a 4402 WLC. A few locations in the building (mostly where the heaviest use is), are reporting connection drops and slowness (connection speed changes several times).

I've been looking at the channel assignments. In one particular wing, for example, the two downstairs APs, which are at opposite ends of the hall, are channels 1 and 6. The floor above is has one access point, in the middle, which is also on channel 6. It's been a while since I've checked, but I think I can get at least some sort of signal upstairs in that area from the APs on the lower floor.

Shouldn't the WLC have assigned channel 11 to one of those? What are the best options here? Should I force one to channel 11?

There's another wing in this same building that's like that, and I'm thinking it might not be as easy to fix. There's 4 APs in a relatively close area. Two are on channel 1, one on 6, and one on 11.

Reply to
pfisterfarm
Loading thread data ...

~ I've got a remote site that has a total of 25 1131 access points under ~ the control of a 4402 WLC. A few locations in the building (mostly ~ where the heaviest use is), are reporting connection drops and ~ slowness (connection speed changes several times). ~ ~ I've been looking at the channel assignments. In one particular wing, ~ for example, the two downstairs APs, which are at opposite ends of the ~ hall, are channels 1 and 6. The floor above is has one access point, ~ in the middle, which is also on channel 6. It's been a while since ~ I've checked, but I think I can get at least some sort of signal ~ upstairs in that area from the APs on the lower floor. ~ ~ Shouldn't the WLC have assigned channel 11 to one of those? What are ~ the best options here? Should I force one to channel 11? ~ ~ There's another wing in this same building that's like that, and I'm ~ thinking it might not be as easy to fix. There's 4 APs in a relatively ~ close area. Two are on channel 1, one on 6, and one on 11.

I don't think I'd worry too much about a channel assignment here or there. Below is my general approach for a situation like this.

First, make sure that there's appropriate coverage throughout the coverage area (with a site survey tool) - for data, make sure that there is at least one AP at -72 dBm (or whatever your standard is), but not too much co-channel interference.

In addition to the site survey (by this I really just mean walking around and seeing what APs you see), I also like to take the "show run-config" output from the WLCs and load it into the WLC Config Analyzer

formatting link
look at co-channel interference, noise, etc. Make sure that things look pretty solid.

Now get a known good client device (like a laptop) and have it run a continuous ping and verify that it can move around everywhere and have good connectivity. (If not, figure out why not.)

Once this is done ... focus on specific clients or specific locations that have more or less replicable problems, and fix them up.

Btw, the Wireless area in Supportforums is quite a good place to ask questions:

formatting link
Cheers,

Aaron

Reply to
Aaron Leonard

Maybe it's worth mentioning that the Windows ping utility can not do a suitable ping. Well, not unless you have a lot of time to wait for a diagnosotic number of packets to be sent:), say an hour or two per sample point.

I use fping.exe from

formatting link
's in the drop down list box) which can do a "cisco style" continuous ping. If you get 100's of packets per second then the the network is working very nicely indeed and you don't have to delay further:) I forget typical values since it's been a while.

I think that the current linux ping does support a "flood" ping but I don't have one handy to check.

I also like to use decent sized packets for wireless surveys, 1400 bytes or so since this stresses the network further.

Reply to
bod43

ping -f [flood ping, root only.] ping -A ['Cisco style', send next ping as soon as response is received.]

Minimum interval in all cases is 200ms for non-root users.

Reply to
alexd

Thanks everyone for the information and links... I have one more question: How do I look at the co-channel interference and noise? Do I need a spectrum analyzer?

Reply to
pfisterfarm

~ On Mar 26, 3:42 pm, Aaron Leonard wrote: ~ > Then look at co-channel interference, noise, etc.  Make sure that ~ > things look pretty solid.

~ Thanks everyone for the information and links... I have one more ~ question: How do I look at the co-channel interference and noise? Do I ~ need a spectrum analyzer?

Co-channel interference refers to multiple 802.11 cells (i.e. BSSIDs AKA APs) overlapping in the same physical space. So as a first cut, you can just use your site survey tool (whatever it may be) and see what APs are present (at what signal levels) at each physical point in your coverage area.

For example, let's say that you see the following APs:

On channel 1: AP1 @ -55dBm, AP2 @ -88, AP3 @ -89 On channel 6: AP4 @ -72dBm, AP5 @ -86 On channel 11: AP6 @ -53dBm, AP7 @ -84

Then I would say that you're fine wrt co-channel - a client in that location should pick either AP1 or AP6, and any other APs on that channel are

30 dB away.

On the other hand, if you have:

Channel 1: AP1 @ -55dBm, AP2 @ -65, AP3 @ -70 Channel 6: AP4 @ -60dBm, AP5 @ -63 Channel 11: AP6 @ -58, AP7 @ -63

Then you have a co-channel problem - no matter whether the client picks AP1, AP4 or AP6, there is another AP within 10 dB.

Now, with co-channel interference, you don't necessarily see any performance problem, until you get under load. To measure the load effect, you really would need a packet capture tool or similar at the time of load.

Noise, on the other hand, is non-802.11 energy. (Or what is *perceived* as non-802.11 energy by the 802.11 radios - it could actually be energy from some distant AP, or from an AP on an overlapping channel [like from channel 3 when you're on channel 1.]) You would use a spectrum analyzer (like Cisco Spectrum Expert) to measure noise.

Having said all of the above ... you can also just look at your lightweight APs to measure the "load" (i.e. 802.11 interference, from this/other APs in your RF group), "interference" (i.e. 802.11 interference from other devices *not* in your RF group) and "noise".

Example (from my lab):

(WLC4402) >show ap auto-rf 802.11b AP1142

Noise Information Noise Profile................................ PASSED Channel 1.................................... -93 dBm Channel 2.................................... -90 dBm Channel 3.................................... -81 dBm Channel 4.................................... -94 dBm Channel 5.................................... -59 dBm Channel 6.................................... -92 dBm Channel 7.................................... -89 dBm Channel 8.................................... -90 dBm Channel 9.................................... -87 dBm Channel 10................................... -90 dBm Channel 11................................... -94 dBm Interference Information Interference Profile......................... FAILED Channel 1.................................... -70 dBm @ 2 % busy Channel 2.................................... -128 dBm @ 0 % busy Channel 3.................................... -46 dBm @ 12 % busy Channel 4.................................... -39 dBm @ 2 % busy Channel 5.................................... -128 dBm @ 0 % busy Channel 6.................................... -128 dBm @ 0 % busy Channel 7.................................... -128 dBm @ 0 % busy Channel 8.................................... -128 dBm @ 0 % busy Channel 9.................................... -56 dBm @ 1 % busy Channel 10................................... -128 dBm @ 0 % busy Channel 11................................... -44 dBm @ 23 % busy Load Information Load Profile................................. PASSED Receive Utilization.......................... 0 % Transmit Utilization......................... 1 % Channel Utilization.......................... 29 % Attached Clients............................. 1 clients [ ... ] Nearby APs AP 00:12:44:b3:e0:00 slot 0.................. -8 dBm on 1 (10.0.47.3) AP 00:19:07:c6:04:b0 slot 0.................. -22 dBm on 11 (10.0.47.11) Radar Information Channel Assignment Information Current Channel Average Energy............... -60 dBm Previous Channel Average Energy.............. -18 dBm Channel Change Count......................... 96 Last Channel Change Time..................... Thu Apr 1 05:44:27 2010 Recommended Best Channel..................... 11 So this shows some real concerns:

- noise at -59 dBm on channel 5

- high interference on channels 1, 3, 4, 9 and 11

- potentially high load (29%) on your channel (11)

- APs excessively nearby (-8 and -22 dBm - they're all in the same rack actually)

- previous channel average energy -18dBm (that's super loud)

- 96 channel changes

If this were a real production network, rather than just in my messed-up lab, I'd be worried ...

Hth,

Aaron

Reply to
Aaron Leonard

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.