The 5 Principles of Elastic WiFi
The 5 Principles of Elastic WiFi
|Elastic WiFi. It’s probably not a term you’ve heard before in relation to wireless and WiFi. I’m going to run through some principles I’ve learnt over a few years in the business, but first, I’ll digress...
Many years ago I learnt to fly. Its kind of similar in complexity to the workload if you were driving two or three cars at once. The plane isn’t glued to the ground, but operates in 3 dimensions. You kick in a bit of rudder to affect a yaw and start dropping a wing in the turn. You can compensate with a bit of aileron to put you in a straight turn – there’s even a little ball in the dashboard so you can check you’re not skidding around a turn.
For every action, there is an equal and opposite reaction and wireless is the same. Let's delve into those principles and see if you can spot what I mean.
In a nutshell, you can not have it all with WiFi. You have to compromise.
1. Less APs = Higher Power
It's simple. You want to spend less, so you buy less APs.
You turn them on and then they power up to maximum because the AP desnity is low – your survey is going to look ok when you do a post-install, but clients won’t roam properly and low power devices will drop out.
With less APs, you will only achieve a data-only grade WiFi and it probably won’t allow low power clients to roam very well.
Lesson: Spend your budget how you will – but get the best advice on what you’ll get for it
2. Too many APs = Higher Channel Utilisation
At the opposite end of the spectrum is the whack-a-mole theory on RF design.
Carpet bomb the place with as many APs and it can’t fail to work properly.
Unfortunately, this isn’t true either. If you skimp on a proper RF survey in favour of adding as many APs as you can, you’ll end up with loads of beacons in the air from all those APs and a high channel utilisation.
Instead of giving you the gold plated WiFi network you expected, you ended up with too many APs and a network that’s like a stretched elastic band.
Lesson: Get a proper RF survey that takes into account the services and clients you actually want to support. It will be cheaper than all those extra APs.
3. High Power = Poor Cell Distribution
So you didn’t put enough APs in, but you’ll fix it by just turning the power up.
Rookie mistake this one - I’ve seen this one too many times.
You will now get big cells from each AP which overlap each other and cause you no end of grief with clients not roaming, high co-channel and generally a poor user experience.
Lesson: You can’t get blood out of a stone. Not enough APs is just that – it's not enough and turning up the power will only help you if you only have high powered clients.
4. Cutting Edge = Bugs
Sorry, but it seems to be a fact of life with cutting edge technology, especially WiFi.
New code is crafted and put into production too quickly and you get burnt. I am sure a lot of people relate to this one.
You get a major new version of code, a new AP type and decide to try some of those fancy new features. . and you’re in a world of hurt. Next thing you’re looking at complex, time consuming bugs, air sniffs and extended periods before you get the network properly operational.
Its been said many times that you need to be careful if you’re on the bleeding edge.
Lesson: Test anything new on a friendly, small, contained site. Get the feedback that everything works as expected – don’t bet the farm on
something new just working.
5. Cheap Price = Poor Design
My personal favourite. You’ve accepted a low ball offer on the cheapest RF survey – it’s a race to the bottom to see who can offer the lowest price. It’s a commodity right? Everyone will do the same job? They’re all trained the same, have the same oversight, experience and double checks on the output?
Don’t kid yourself.
The cheapest price will involve some corner cutting – and that corner cutting is going to cost an awful lot more to remediate down the track once you have to revise, add more APs and move the existing ones around.
Lesson: Don’t mix up cheapest price versus best value. The cheapest quote probably won’t deliver you the best outcome. Do your homework – if they’re not asking you how you intend to use the network, they probably are going to do a cookie cutter with the most inexperienced person and deliver you something that’s an approximation of what you actually wanted.
Engineers are people and they are all different – the best qualified command higher prices, so bear that in mind when you select the vendor of choice.