When looking at wireless access points, I would always find myself curious about one thing:
How many concurrent users can use this thing at the same time, before it falls over?
It’s a question for which the answer isn’t plastered on the side of the box. More often than not, you’ll find asking one of the technical teams at the manufacturer of your choice will also reward you with vague uncertainty or general disinclination to answer the question.
On the odd occasion where you may find your question being answered, it could come in a variety of quantities:
“10!” … “50!” … “100!” … “200!”
And then you get to the caveats…
“…depending on usage…”
To understand the proper answer, you have to understand the nature of the question itself. And furthermore, understand the logic behind the technology.
Let’s roll-back to a few years ago, when 802.11n was still an idea on a chalkboard, and 802.11g was the cream of the crop. Routers on every shelf in every computer store boasted of “Super Fast” 54Mbps speeds.
The way that these speeds are measured means that in actual fact the bandwidth a single user would get will be barely half of that. Practical demonstrations of 802.11g routers delivered throughput with full signal at around 19-22Mbps. This is when tested with a laptop/wireless device with a similar 802.11g card in them.
So, 22Mbps we’ll use as the example for our pie. For simplicity’s sake, we’ll say every user connecting uses 802.11g for connecting. Your first user gets the whole pie in one piece, if a second user comes along and connects that pie is split equally between the two. If a third arrives, it splits into thirds, and so on.
Going by that method, a base starting bandwidth of 22Mbps could theoretically support approx 20 users that want to use a service at a (by today’s standards) slow speed of approx 1 Mbps. And if you add any more on, they only get a smaller slice of pie each.
Now we fast forward, and come into the age of 802.11n (and soon to come, 802.11ac!). Whilst the speeds are greater, the logic behind them stays the same. It’s just a bigger starting pie for everyone to tuck into.
And it isn’t really something that can be beaten by throwing money at it. You could have a store-bought £50 54Mbps wireless router, and a £300 54Mbps Enterprise router, and have the same throughput (trust me, I’ve tested such things, and was surprised by that revelation myself!). All you get is the extra bells and whistles (additional monitoring, intrusion detection, multi-SSID, 3/4G backup, etc.). Whilst this might be good in some instances, it doesn’t really bolster the number of concurrent connections. Only bandwidth and radios can achieve that, really.
So now you know about the type of numbers your chosen access point can support, you find that one on its own just won’t cut the mustard. But if you put two access points on a shelf next to each other, you won’t be really much better off.
In comes a new player to our game; cross-channel interference.
For 2.4 Ghz Wi-Fi operation, the wireless spectrum comes neatly packaged in 14 “channel” blocks. These are 5 MHz apart from each other. However, wireless transmissions themselves occupy 22Mhz of channel bandwidth. If we were to use channel 1 (2412 Mhz) for one access point, and channel 2 (2417 Mhz) for another (within the same area), we would find degraded signal quality/throughput as both battle for airspace in that cross-over space.
For that reason, there are really only three channels which play nicely with each other. These are channels 1 (2412 Mhz), 6 (2437 Mhz), and 11 (2462 Mhz).
So, whilst you have up to 14 channels (in some countries) to choose from, you only really have 3 that are of any practical use in a properly planned/deployed wireless network.
This is where brick walls and thick concrete flooring and ceilings can be your best friend. The dBm attenuation (signal loss) caused by these thick solid obstacles mean that you can “isolate” floors and rooms by using the environmental challenges for your own advantage. Challenges which, ironically, would be something that you were trying to get around in previous network plans.
However, it isn’t all doom and gloom – there is a better long-term solution to your crowded wireless networks. It’s called 5Ghz.
Whilst 2.4Ghz wireless channels ended up having a rather frustrating level of channel overlapping, 5Ghz channels provide a much more convenient channel distribution and subsequently eradicates the issue of channel overlap between your access points.
This suddenly presents a network planner with the advantageous route of deploying more AP’s on 5 Ghz in a much more tightly-packed space. This provides more access for more users, whilst also avoiding the interference which plagues older 2.4 Ghz networks.
The only drawback with 5Ghz, is that it isn’t as widely adopted with regards to wireless devices. As time goes by, more devices are coming out with “dual radio” wireless cards, allowing users to switch between 2.4 and 5Ghz networks as they please. In a public Wi-Fi hotspot setting, you need to look towards accommodating both network types, as you cannot guarantee that all your visitors will be compatible with 5Ghz frequencies.
However, if you are planning for a more office-centric environment, then you could look towards ensuring any wireless devices (and their subsequent compatibility) are 5Ghz compatible to make a far more future-proofed network. Such concepts as BYOD (Bring Your Own Device) could be implemented in accordance with this so that you define your own operating policy within your company (enforcing 5Ghz preference, and eradicating use of 802.11b on the network, etc.).
So in conclusion, always remain mindful of how many connections you want to handle for a wireless implementation, and the environment you’re deploying in, to ensure you make the most-informed choice that you can for optimal results.