Wireless Access

I have a number of outdoor antennas - AP275's and down the hill in an enclosed area a 225. It's a zoo, so these AP's are named Red Panda Deck, Otterly Cool, Naked Mole Rat and Panda Veranda.

Nearest to Red Panda Deck is a Foscam IP camera.

The camera associates with one of the others, and not Red Panda Deck.

I vaguely understand the RF coverage "lobes" and am assuming that the proximity of the camera to the AP is what's causing the failure to associate.

Short of moving the AP around (not an easy proposition), is there something else I can try? I see in the controller Antenna bearing and tilt angle - is this something I can play with? I don't see where I can manually change radio strength on these, and am not sure what I would do with that anyway.

The AP is mounted to a pole and is about 11' off the ground and is close to the same height as the camera, which is 20' away. I can't move the AP higher, but I could lower it by a foot.

Thanks!

The antenna bearing in the controller (in the AP provisioning window) is just for recording the physical alignment, it has no actual control or affect to the actual.

It would be helpful if you had a drawing whoing which model of AP is where, what antennas are where and pointed in what direction, etc. but the Foscam, if wireless, would pick whichever signal it sees as better. You could go down there with a wlan survey tool (InSSIDer or something like it) to check at the Foscam camera what BSSIDs are the strongest. Correlate that to which AP is where in your examples above and see if it makes sense (using a 'show ap bss-table' will give you the AP BSSIDs). 

However, it sounds like the Red Panda Deck AP and camera are close enough that it should work. All these APs have the exact same config, are in the same group, etc? If so, as a test, you could disable the wireless SSID on all but the red panda deck AP and see if that forces the camera to attach and if not, then there's some kind of L1 problem between those two devices you would have to troubleshoot.

Thanks. Yes, the 275's are all in the same group with the same configuration. I'll see about sketching this, but to describe in words:

Using Red Panda as origin point:

Naked Mole Rat is  SE125' away with an intervening building and a bamboo grove

Otterly Cool is NE 125' away with an intervening couple of buildings

Panda Veranda is West, very much downhill, and 100' away under a roof.

Today, I have these crazy associations and signal stength

10:43 AM Otterly Cool -
10:33 AM Red Panda Deck -
10:23 AM Otterly Cool -
10:13 AM Naked Mole Rat South -
9:33 AM Otterly Cool 8
9:13 AM Naked Mole Rat South 11
8:53 AM Otterly Cool 19
8:43 AM Panda Overlook 0
8:13 AM Otterly Cool 22
8:03 AM Panda Overlook 0
7:53 AM Red Panda Deck 0
7:33 AM Red Panda Deck 0
7:23 AM Otterly Cool 17
7:13 AM Naked Mole Rat South 0
7:10 AM Panda Veranda-AC -
6:53 AM Naked Mole Rat South 23
6:13 AM Red Panda Deck 36
5:53 AM Otterly Cool 20
5:33 AM Outback South 7
5:23 AM Panda Veranda-AC 0

But no match events for today. The only ones I have are:

11/12/2014 3:16 PMNaked Mole Rat SouthRed Panda DeckSticky Client

11/12/2014 3:04 PMNaked Mole Rat SouthRed Panda DeckSticky Client

And here's a picture to highlight my confusion. I'm beginning to think this AP is defective. Why would it briefly connect with a signal strength of 36 and then associate with other, far more distant APs? And those associations are weak.

It's not the camera. It worked flawlessly when inside, hitting Aruba 105's.

And as you can see, the camera is line of sight with the AP, at the same height, and ~25 feet away.

The controller doesn't have any complaints about the AP. The radio strength:

doesn't seem out of whack.

This may be caused / aggravated by the AP bring 802.11ac and the camera being something else.  I notice you are using 80MHz channels.  (149E).

There could be two different issues here.   One is that the camera if it is older 11n or 11a/b/g camera can only communciate at legacy rates and channel bandwidths (20 or 40MHz).

The other is that there is generally a lot of interference in outdoor environments, and by running 80MHz channels the APs are exposed to FAR more interference than in the past.

In general, the best practice outdoors is to run 20MHz channels unless you have some specific reason to run wider channels.   And in that case you need to get out a spectrum analyzer and make sure the air can support your intended deployment.

So I would take these APs to 11n, 20MHz  (HT20 mode) and then see how things work.

Hmm. Well thanks, but the whole idea behind the 275 was to offer AC and some measure of future proofing. The camera is being steered to more distant, identically configured 275's. At the moment it is associated with Otters, which has bamboo and buildings and distance going against it. While I could experimentally chop the AC channel on this one AP, I certainly wouldn't want to do that throughout. It's cold today and the place is empty. My landscape is a bowl, and there is no outside interference. The only thing in my RF spectrum is Aruba AP's. So I don't really see how it can be spectrum interference,

It may be the camera's radio (although if it's b/g/n I would expect it to associate with the correct radio on the AP). Because otherwise I have 33 of these AP-275's serving the zoo really, really well, both with newer AC devices as well as the legacy 802.11n devices.

One thought occured to me. Can I create an SSID that only associates with one AP? So if I had an SSID of RPCAM and only broadcast it from the Red Panda Deck, then I could set the camera to be on this SSID and see what goes. Can I do that? If so, how?

camera is only up to 802.11n

its current connection is:

 11n 2.4 GHz (20)

but that connection is to a more distant AP and the SNR is 20. And weirdly, the health of the connection is 100%!

80MHz channels are never recommended in high density areas, indoors or outdoors.  Are you providing free wifi to guests?   You will still be providing 11ac even in VHT20 mode (e.g. higher data rates).  Just not the highest possible.

In high density areas you want to provide more small channels in more places, not fewer big channels in more places.  Because - as you are finding - RF travels a really long way and so none of these APs really is isolated from the others.

But back to your specific issue.  Your last comment re-2.4GHz suggests the client doesn't like 5GHz for some reason.

Have you verified that the camera is using the same channels as the APs?    The camera probably treats channels 36-48 as usable for instance, while the AP275 will not because those channels cannot legally be used outdoors in USA and 275 is an outdoor product.   That could be producing a mismatch.

Since we don't have DFS approvals yet for AP275 you won't have a mismatch on channels 100-140, but this is something to look out or in future code upgrades when we turn this on.  (e.g. if the camera cannot do channels 100-140 then it may not see the AP).

Yes you can create per-AP SSIDs.   Easiest way is to create separate AP groups for each AP with custom VAP profiles.   There is a way to do with it AP-specific config but that is tedius.  TAC can help you either way.

The camera should be connecting on 2.4GHz to the closest AP even if it cannot see the 5GHZ.  You may need a protocol analyzer or spectrum analyzer to figure it out.

Even if you are in a bowl, you may have other unlicensed systems in 5GHz.

-cl

thanks for your patience with me.

I noticed a high number of these snmp messages:

wlsxNRadioAttributesChanged wlsxTrapAPMacAddress.0: AC:A3:1E:C0:02:14,
wlsxTrapAPChannel.0: 11, wlsxTrapAPTxPower.0: 24, wlsxTrapAPIpAddress.0:
10.1.2.74, wlsxTrapAPRadioNumber.0: 2, wlsxTrapTime: 11/13/2014 15:54:54
UTC-5

I get 5 of them all in the same minute, and then again every couple of minutes.

Thanks, did not know that about VHT20 mode and 11ac. There's quite a lot I don't know about this. It's PFM.

I don't know that I'd characterize this as high density (yet), but I'll look into the settings. Does VHT20 *not* use 5ghz? I don't much care about providing my guests with the highest possible speed, but would like to offer them consistent service. I thought that 5 ghz had less distance to it and less congestion in the channels, both of which seemed like a good thing.

I've had the current setup operating for a few months. I have had a max of about 30 clients per radio. I haven't received any complaints, and my own experience is great with an iPhone 6, an older chromebook, and older PCs. I've had some dropcams running off of other AP275s no problem.

Anyway, thanks for the help and I'll start a conversation with TAC about the mode.

Still utterly confused:

and interestingly, all of those association times are separated by a factor of 10 minutes.

Hmm, looks like variations are in multiples of 10 exactly minutes.  some 10, 20 some 40 and one 60.

Feels like the client is doing something to me.  Only way we would be forcing him off is if the user idle timeout kicked in, but that is not 10 minutes and your device should not be idle.

I see some assocations to the close AP at 50 SNR.

The SNMP trap you listed could be a normal ARM channel change due to interference.   ARM should not normally change channels with any client associated to it  (e.g. ClientAware is set to enabled by default).  This is unlikely to be the source of the reassociations since ARM doesn't work on timers but rather based on metrics that build up over time.  You could attempt to correlate your SNMP messages with your roaming lots to confirm this.  

My gut from what you're describing is that the wireless radio and/or antenna in the camera is not performing well.   Is it possible to connect a wireless client bridge to the camera ethernet port and monitor for a difference in behavior?  

Suggest you work with TAC on this.