Friday Oct 24
Jan
19/12
Why more APs aren’t always better
Updated on Thursday, 9 February 2012 12:42
Share

Lately, I’ve been forced to dispel a volume of wireless myths, both in way of technology and vendors. I’m not sure if it’s a full moon, or some other astrological occurrence, but it’s gotten a little crazy recently. So, I thought I’d take a few blog posts to address some wireless myths, in brief, to keep in the back of your head as you explore wireless solutions and upgrades in your environment.

Here’s a thought on why more APs aren’t always better.

The other day someone said “that’s fine, if that happens, we can just add more APs, they’re pretty cheap.” As I looked over the predictive wireless survey (APs and predicted coverage mapped out on a floor plan), I realized there were already quite a few APs in the general area he was referring to. “Well, here’s what we need to watch out for…” I replied, and here’s what we discussed…

More APs isn’t necessarily better. For this to make sense, we first need a very basic understanding of the duplex and collision avoidance properties of wireless (see Understanding collisions and duplex in wireless). With that overview, we learn that that wireless things can’t listen and talk at the same time, and that multiple wireless things can’t transmit at the same time, because that results in collisions.

If you could visualize RF in the air, and see the little packet-carrying waves passing by, you’d start to realize that everything wireless is talking, and all at different intervals. Most notably, always-on wireless devices tend to be a bit chatty, whereas laptops, tablets, Kindles and other things running on battery tend to hibernate a bit, and send only occasional messages when they’re not in use.

The actual APs, however, are constantly sending messages. They’re sending management messages to clients, coordinating air time with request-to-send and clear-to-send packets. They’re broadcasting beacons and wireless network information so new people can connect. Some may be negotiating channel and power settings. All-in-all, they’re filling the air with little tidbids of AP-talk all day, every day. Knowing this, you can probably start to see why more APs in a given area might not be beneficial.

With too many APs constantly talking in a single area, we start to infringe on the time that clients can use to transmit and receive wireless data.

If we go back to our analogy of the classroom and the professor (from  Understanding collisions and duplex in wireless), we can modify it slightly pretend we have 10 professors in the room; maybe we’re having a panel discussion to re-evaluate Einstein’s principles of general relativity. If the professors all start talking a lot, then that leaves less time for the students to interact and engage. To better prove our point, we’ll say the professors are native German-speaking fellows, so when conversing with one another, they switch to German to facilitate faster conversation. The students are all English-speaking, so they’re not benefiting from anything the professors are discussing amongst themselves; they just sit back, raise their hands, and wait for the chatter to die down so they can speak. If you’re measuring the success of the class by how many questions get asked and answered (think wireless bandwidth from clients) then too many chatty professors makes for a fruitless day.

In a wireless environment, a similar occurrence happens when too many APs are in one area. As a rule of thumb, we generally say at any given point in space, if a client were there, it should only be able to see a maximum of 5 APs. Two to four is more normal; five is on the high end of the spectrum, but may be needed in the case of traditional microcell solutions that require channel staggering and AP density to support client density.

Each vendor will have a general rule of thumb for AP counts also, based on unobstructed square footage. For example, a vendor may generalize that they recommend one AP per 2,000 sq ft. Keep in mind, these are always just extremely rough starting points, because actual AP count and placement depends on user density, RF environment and building materials. What you won’t (or shouldn’t) see is a vendor recommending extremely high AP density, regardless of how much they may want to sell you ‘extra’ APs.

And, that’s why more APs aren’t always better

# # #


1 Comment
  1. CommentsDevin Akin   |  Friday, 20 January 2012 at 12:59 am

    Awesome.

    Just thought I would share a few thoughts of my own along these same lines:

    http://blogs.aerohive.com/blog/wi-fi-that-wont-die/fewer-aps-is-always-better

    Hope it helps.

    Devin Akin
    Chief Wi-Fi Architect
    Aerohive Networks


Leave a Reply