Wi-Fi Speeds-n-Feeds Are So Old-school


Speed-n-feeds are not the future of enterprise Wi-Fi. Speed-n-feeds are like my grandmother’s potatoes. I’ll explain.


Speed is a given. Speed is a commodity. This is what you talk about when you have nothing else to offer, such as system intelligence. Some companies keep trying to rehash the speeds-n-feeds story like my grandmother used to treat potatoes. First, you’d have baked potatoes. If you didn’t eat all of them, the next night, you’d have mashed potatoes – made from those same potatoes, of course. If there happened to be any left-overs after that, you’d have fried potato cakes the next night. Believe me, the list of how those potatoes could be served was endless until those potatoes were gone. Same potatoes, different day.

With 802.11ac, networks have plenty of speed. In fact, I will bet you a cold beer that far more bandwidth goes unused on a daily, hourly, and per-minute basis than is ever used – even with an 802.11n network. You would be hard-pressed to find an AP in any enterprise environment that exceeds an average of 10% of its throughput capability when averaged over a normal 8-hour work day. If you find even one, please give me a shout with statistical proof. I’m interested to see that data.

There is Wi-Fi bandwidth to spare within most enterprise networks that have 802.11n or 802.11ac technology. The bottleneck isn’t in the radio capability, the architecture, the channel airtime, the AP CPU power, or even the available spectrum. The only exceptions are major interference sources, which can affect channel airtime and available spectrum, but that’s off-topic because you don’t design capacity around interference sources.

If there’s an actual speed bottleneck anywhere, it’s found in the battery life of mobile devices, which causes the use of single spatial stream (1SS) radios. Truth be told, even mobile devices using 1SS are fast, ranging from 65 to 433 Mbps, with about half of this data rate being considered usable throughput.

If you have some kind of crazy high-density scenario, sure, you could potentially run into an airtime bottleneck on specific APs at specific times, but that’s a sub-one percent use case issue in most enterprises, and buying a entire solution around a sub-one percent use case seems silly given the price multiple that you are apt to pay.

Most vendors are still building wave-1 802.11ac APs, and the industry is already talking about forthcoming wave-2 technologies, which are more than a year away, highlighting the marketing frenzy around performance. Look, I’m not against performance, but I’m saying that it’s far from the most important consideration.  In fact, at this point, it’s pretty far down the list.

Well, since there’s plenty of speed, what aspects are more important?

  • Solutions focused on vertical markets
  • MSP (managed service provider) enablement
  • Simplicity and intuitiveness of use and deployment
  • Sales model and process: capex, opex
  • System security, redundancy and stability
  • System maintenance, monitoring and compliance reporting

Those are just a few, but they are enough to point out that all of the hoopla around “just speed” is off-target. Almost any enterprise vendor can now provide reasonable connectivity, and they have the reference customers to prove it. Therefore, winning in the market is no longer just about connectivity, but rather about solving customer problems by providing a focused solution.

You like apples? 😉

/Image via foodieandfellow.com. For an uncommon twist on potato cakes, get the recipe: Orange ricotta sweet potato pancakes

Devin Akin

Devin Akin



  1. GT Hill says

    I agree that there is plenty of speed for now but, repeating from my own blog, there is a HUGE difference among enterprise APs can handle with regards to active associations. I can’t comment on AT APs since I’ve never tested them but among the tested APs out there the difference is significant.


    • says

      Hi GT,

      Thanks for your comment. I agree that there is a difference in the area of active associations, less-so when you’re using encryption. However, it’s my opinion that it’s really not the number of active associations that matters so much, but rather the number of users/devices that can achieve their targeted application throughput in a given environment. It’s all environmentally-dependent, which is, in-turn, vertical-market dependent. For example, a classroom may need to support 30 1SS devices with 2Mbps stream video, while a quick service restaurant (QSR) may only need to support 20 devices with web browsing and email with a collective Internet backhaul of <10Mbps. Thanks again for dropping by.

  2. says

    I also agree that for most uses, there is plenty of speed on the AP back haul. i deal mainly with schools and we tested lots of wireless systems and found for basics most worked. but when we drilled down into actual use cases such as 30 logins with large profiles or other school tasks we found that not many worked as expected.

  3. says

    Hi Jacob,

    Thanks for your comment. You mentioned logins with large profiles. Does this mean that you are using Windows laptops in the classrooms? Most people are using or are moving to iPads, Chromebooks, and similar devices in classrooms these days.

    Additionally, classrooms in EDU is just a single use case. There are a large number of other vertical markets, e.g. Retail, Healthcare, Financial Services, State & Local Gov’t, Federal Gov’t & Military, Manufacturing, Logistics, Transportation, etc where no such use cases exist. What are your thoughts on such?



  4. says

    Hi Devin,

    Yes it does mean windows devices, we have also seen a sharp increase in the number of BYOD type devices such as the IPADs. my thoughts on other use cases such as office or factories tend to be more about the RF setup and how well APs change there channels to meet wifi needs and also how they load balance traffic.

    The best part of AC for me, is not here yet which will be spatial stream multiplexing. as most devices are not using even 2×2 in most use cases. also beamforming features and client features i feel will become better used across most vendors in next 12 months or so.

    What do you think on AC would you wait till the above features are added?


  5. says

    Hi Jacob,

    Thanks for your reply.

    I can definitely see how office environments and factories would need good RF handling, but not sure about load balancing. I will concede that load balancing can be important in some scenarios, as it depends on client density and channel utilization.

    For 802.11ac, I assume you meant MU-MIMO. To answer your question, no, I wouldn’t wait. I would begin deploying 802.11ac as it makes sense, but I would keep in mind that 802.11n is still useful and 802.11ac Wave-2 will be able to handle even higher density (especially when there are many single-spatial-stream (1SS) capable devices in the network). I think Wi-Fi upgrades are generally “rolling upgrades”, meaning that they are pretty much always happening because the technology itself is always being upgraded. With all of that said, I would only buy what you NEED, keeping in mind that 802.11n is still very fast in most scenarios.

    Hope this helps,


  6. says

    Hi Devin,

    I am not advising customers to wait, I just wish the features were here already. On a side note how to do you feel Airtight compares to the other major WiFi vendors in the AC market?


Leave a Reply

Your email address will not be published. Required fields are marked *