Are You(r APs' Transmit Power) Still Down? Raise 'Em Up

Remember me? (lyrics NSFW)


Too many enterprise Wi-Fi deployments experience Wasted Wi-Fi.  

Wasted Wi-Fi is what happens when an enterprise wireless network offers users lower performance than it would under an optimal configuration.

Today, the most common culprit of Wasted Wi-Fi is low AP transmit power.  So, if your AP transmit power config is Still Down, Raise 'Em Up.

I've noticed a lot of Wasted Wi-Fi recently.  Situations where, given the distance, client density, and physical environment, the Wi-Fi should be faster.  These are situations where, irrespective of issues that may exist in the infrastructure equipment or design of the wireless LAN, client devices are getting worse performance than they should.

By its definition, Wasted Wi-Fi lays entirely at the feet of those who configure enterprise Wi-Fi settings.  The good news is, that means Wasted Wi-Fi is solvable without expensive and time-consuming changes to the wireless LAN infrastructure.  The bad news is, solving Wasted Wi-Fi takes guts.

Changing configurations on a large scale Wi-Fi network can be a harrowing experience, especially if the organization is not beset with complaints about Wi-Fi performance.  In fact, a compelling argument can be made that configuration changes should not happen under such circumstances.  "If it ain't broke, don't fix it," the old axiom goes.

Long time readers of this here blog may recall that I don't believe in reactive Wi-Fi optimization.  That's because I hate failures.

Wi-Fi failures can damage the reputation of an organization, and often go unreported.  For example, the person who got me into the Wi-Fi business lo these many years ago recently told me about chronically poor Wi-Fi at the San Manuel Hotel & Casino in southern California.  He said "everybody knows" that the guest Wi-Fi in the casino is "unusable", and has been for years.  My pal even said that he will occasionally drive to a further away casino, Pechanga, simply to avoid the frustration of trying to use Wi-Fi at San Manuel.  "Have you ever complained?" I asked?  Of course he hadn't.  He's at the casino for a few hours, thus he doesn't want to waste his time dealing with tech support.

Avoiding Wasted Wi-Fi can prevent failures by making Wi-Fi channels as efficient as possible.  Efficiency in Wi-Fi is defined as getting the most data across a Wi-Fi channel in the least amount of time.  It is important because time is constant.  One cannot stretch or compress a period of time.  Ten seconds is always ten seconds.  Skilled Wi-Fi professionals can, however, stretch and compress variables that can allow more data to flow over a Wi-Fi channel during that ten seconds.

In my travels, I have noticed that the top cause of Wasted Wi-Fi is low transmit power.

There are other causes, to be sure.  Some Wi-Fi networks use narrow, 20 MHz wide channels, when wider, 40 or 80 MHz wide channels would improve performance (or use wider channels, when narrower channels would increase stability).  802.11r Fast Transition (FT) is needlessly disabled on many enterprise Wi-Fi networks, causing network hiccups for mobile client devices.  Co-channel interference (CCI) can also be a Wasted Wi-Fi villain, even in the 5 GHz frequency band.

Low transmit power deserves special attention because its effects seem to be misunderstood.  People seem to understand the pros & cons of channel width, FT and CCI.  I find that many people who oversee enterprise Wi-Fi settings see transmit power only as a method of increasing or reducing "cell size", and thus are unaware of its potential to cause Wasted Wi-Fi.

When AP transmit power is set too low, Wasted Wi-Fi can happen in two different ways: low downlink data rates or high downlink Retry percentages.  Some APs will reduce data rates if AP transmit power is too low, while some APs will stubbornly attempt to use higher data rates, and live with an increased percentage of Retries as a result.

The manifestation of Wasted Wi-Fi -- at least, based on what I've seen -- varies by vendor.  For example, I have found that Cisco APs tend show to high downlink Retry percentages when Wasted Wi-Fi is happening, while Aruba APs tend to show low downlink data rates.  I am a big believer in the old Russian proverb, "trust... but verify", so I encourage everyone to do some monitor mode captures and test their own APs.

If you are open to adjusting AP transmit power and you want to check whether Wasted Wi-Fi is an issue, I recommend using a symmetrical application.  Facetime was my app of choice.  I used it in the following example.

I began a Facetime video call from a location where my iPhone was getting about -65 dBm signal strength (RSSI) from the AP.  After running the capture for a few minutes, I applied the following filter to isolate uplink Facetime traffic:
wlan.ta == [My iPhone's Wi-Fi MAC address] and wlan.fc.type_subtype == 0x28
I saw the following:


Data rates varied over the entire time of the Facetime call, but at and around captured frame number 188700, the client-to-AP data rate was 117 Mbps.

I then applied the filter for looking at downlink Facetime traffic:
wlan.ra == [My iPhone's Wi-Fi MAC address] and wlan.fc.type_subtype == 0x28
The result was lower data rates, with 87 Mbps being used for AP-to-client data during the same period of time:


During this time, uplink and downlink Retry percentages were both in the 4% range.  That made it clear that any significant discrepancies in uplink/downlink data rates were likely due to Wasted Wi-Fi.

In this instance, instead of getting on and off the channel at the speed of a 117 Mbps frame, the MPDU was taking 35% more time.  That's hundreds, thousands or potentially millions of frames needlessly eating up more channel time than they need to. 

To be fair, my Facetime call worked great in the low user density environment in which I tested.  But this is about preventing failures.  When that same location endures times when it is lousy with Wi-Fi users, the Wi-Fi tends to break down.  (There was a reason I was doing the testing, after all.)

The Wasted Wi-Fi I was seeing made sense once I learned of the AP transmit power configuration for the network.  Aruba AirMatch was being used, with a an AP conducted power range ("conducted power" is the term Aruba uses for transmit power) of 6 to 9 dBm.  (The actual configuration was for 15 to 18 dBm, but Aruba allows for configuration of EIRP, not conducted power.  For the APs I tested, I had to subtract 9 dB from EIRP to determine conducted power.)  Even the maximum AP transmit power configured on this network -- 9 dBm -- was far lower than typical client transmit power.

What was happening on this network was my iPhone was using a higher transmit power than 9 dBm, and the client had settled into 117 Mbps as a viable data rate.  My iPhone's AP had its transmit power ("conducted power", in Aruba parlance) maxed out at 9 dBm, so it had to settled into a lower data rate in order to avoid chronic Retries.  (Cisco APs, in my experience, tend to use data rates that mirror or exceed client data rates, then suffer through excess Retries when low AP transmit power is configured.)

My overall recommendation is to configure AP transmit power to a similar level to client transmit power.  Different clients have different Wi-Fi behaviors, but as a general rule I have found that a transmit power somewhere between 12 to 17 dBm works well.  Raise your AP transmit power up to that range, and you will give yourself the best chance of avoiding failures due to Wasted Wi-Fi.
***


If you like my blog, you can support it by shopping through my Amazon link, becoming a Patron on Patreon, or by donating bitcoin to 1NT4CwZZFMxKxrdY7txbfKujFNHzD7tezA



Thank you.

Twitter: @benmiller

ben_miller at icloud dot com





Comments

  1. Thanks Ben,
    I totally agree with everything you said - I just never got into the packet captures like you did. Very glad to see you took the time to prove it out. I've been using 17dBm for a while now as my standard target Tx Power but I never considered giving it a range of 12-17, thanks for that insight!

    ReplyDelete

Post a Comment

Popular posts from this blog

Go To Sleep, Go To Sleep, Go To Sleep Little iPhone

Roam, If You Want to (As Long as Your Channels Are 20 Mhz Wide)

WPA3 Adds Four Security Enhancements, One of Which Matters