Wireless Access

The whitepapers on Aruba ARM would have me believe that ARM is so great that I can deploy a wi-fi network without tweaking the RF/ARM profiles.  For instance, I shouldn't need to worry about setting min/max EIRP values because ARM will adjust power levels dynamically based on the RF environment.  However, I know this can't be true as my fellow peers often talk about tweaking these values.  The primary reason I ask is because I recently did my first Aruba deployment and I'm not sure if I should've tweaked the EIRP or let ARM do its thing.

 

So how do you know when the EIRP values should be changed?

 

Does the need for tweaking EIRP indicate a failure on ARMs part?

More often than not, it is when you have too much co-channel interference.  This is mostly on the 2.4 Ghz radio and can be because of a number of factors like:

 

- density of APs

- interfering/rogue APs in the area

- non 802.11 interfering devices like cordless phones and microwaves

- vertical placement of the APs floor to floor

 

If you look at the APs out there now and a certain area (or all of them) are set to the min Tx value, that is telling you that there is enough reason to use the lowest possible power setting on that radio.  This may or may not be 100% clean in your network.  A good report/chart you can reference in Airwave is RF capacity and RF health report.  This will tell you the channel utilization that is being seen.  The controller will also show this in the dashboards but it is mostly "real-time" and doesn't keep a long history of this data.  During normal client activity, it is acceptable to see spikes in channel utilization, however, if you are seeing a consistent high percentage even overnight or without users, than you proabably are suffering with co-channel interference.  

 

By lowering the Tx value from 12 or 9 to 6 as an example, will tell the APs that IF you need to lower your power from the current value to a lower number, we will allow it to happen.  Keep in mind that this setting is done PER AP GROUP. If you have a mixed environment and you are seeing high channel utilization in only a few environments, then my advise would be to either break off the APs into a new group or just modify the group seeing the issue with this new ARM profile.

 

Other than reducing the min Tx value, there are other settings that you can modify to alleviate this should this be the issue.  

 

1. You can drop all broadcasts/multicasts in the air.  This will reduce the amount of packets that many consider to be unnecessary in a shared wifi network

2. You can enable "deny broadcast probe".  This prevents a client from sending a broadcast SSID discover packet on the network and EACH AP that can hear it, must respond.

3. You can enable band steering to reduce utliization on the 2.4 and move capable clients to the 5 radio

4. You can disable the low transmit and basic rates of the 2.4 bands.

5. You can set a local probel request threshold so that associating clients must have a minimum SNR value.

 

A discussion with your SE, partner, or TAC should help put you on the right path as to what to do with the RF.  ARM is a great technology but it isn't a 100% set it and forget it.  However, once those initial settings and tweaks are done, you should be in for optimized RF and happy users!

Thank you, Seth.  I had forgotten about checking the APs to see if they're all operating at the min TX power.  That had been suggested to me in the past, and I had forgotten.  Appreciate the refresher.

 

I'm doing 4 out of the 5 things you've suggested.  The only one I haven't messed with is "deny broadcast probe".

 

I was expecting a response like "yeah, you really should've enabled ARM for XXXXX reason".  I'm glad to see that isn't necessarily the case.  I'll keep an eye on the power levels and lower them if needed per your advice.

What are the ramifications on disabling the low transmit and basic rates of the 2.4 bands? I thought I read somewhere that by turning those off you may alienate some clients that work by establing a connection at a lower rate and then ramping up to the appropriate rate. Do you see any problems with disabling those?

 

Thanks!

Pruning the lower transmit/basic rates frees up air time as the radio is able to transmit at higher rate on average.  Typically, you'd prune off the 802.11b data rates since a lot of environments don't have legacy 11b devices any more.  As you pointed out, there are some b/g devices that want to see the lower basic rates advertised in beacons, or they will not associate to the network.  The example I hear most often is the Nintendo Wii.  I have disabled the lower transmit/basic rates on my 2.4 radios and saw a dramatic decrease in utilization.  I would highly recommend doing so if you know for sure that you don't have legacy clients in your network or devices that need to see the lower rates.

 

I made a post on my experience with pruning the 802.11b rates: http://community.arubanetworks.com/t5/Unified-Wired-Wireless-Access/2-4GHz-Unusable/m-p/72790#M2315

If you look above I mentioned doing this only for a certain AP group. You typically would not want to remove the lower transmit rate in say a warehouse that may have legacy Motorola scanners.

Sent from my iPhone

 

In my experience , I usually make changes to this values when you have a very High Dense Deployment of APs primarily on the 802.11g band since there's not that many channels available.

 

ARM works really well and assigning power/channels but some cases where you have a lot APs (802.11a Dense Deployment) some the 802.11g radios might hear each other too much so may want to lower the min EIRP value so that way ARM can make the approatiate adjustments 

 

We decided to split the ARM settings into .11a and .11g  that way when you makes changes to these values you are only impacting one band

 

 

 

I know this is an old thread, but given that it just came up in my search results...obviously people are still hitting it.

 

Most of the replies above focus on adjusting the levels used by ARM down for dense deployments. I actually have the opposite situation in my network. We have some human health monitoring devices which send out a little "heartbeat" every 15 minutes. They have a tiny antenna and run on a AA battery, so their transmit capability is extremely low. If ARM turns down the power settings too low, they can miss a beacon and cause a panic. To alleviate this, we turn up that minimum power setting so the radios are always able to reach out.

That's a good point. In the old hallway approach we often had to turn up the minimum which may have made for a better receive signal but very few clients were capable of making full use of the signal. Thankfully we have moved away from that approach.