Which Access Point Is Fastest?
"There’s something special about experiencing the full speed of fibre for the first time. After coming from my home with a slow 6Mbps fixed wireless line (which, I should add, is not symmetrical, so my upload speed is woefully slow), Vanilla’s fibre network in the office blew my mind. It was unbelievably fast, and I had no idea that plebs like myself and the other interns that had started would just be allowed to use all that speed at whim.
It was, however, not all fun and games. The first part of our internship (after the initiation phase of crimping cables for days) required us to test the wireless speeds of various types of access points. Access points are simply devices that allow you to create a wireless network. We've all been there: you waited way too long to start your assignment and now you only have a few minutes left to submit, when suddenly the WiFi stops working. Now there are 2 things you can do that immediately comes to mind: either you wait for the WiFi to reconnnect (which is no good if your assignment is due in the next few minutes), or you create a hotspot on your smartphone. By doing this, you've turned your phone into an access point. In order to make a hotspot (really just a fancy term for a location with wireless internet access), you need an access point, which is typically easily portable. The access points are connected to a router (or it can be built into the routers themselves), and allow you to connect to a wired network wirelessly.
While not exactly reaching the gigabit speeds we get with a wired setup, the speeds we've found aren’t something to sniff at. Using the 5GHz frequency band, we’ve gotten symmetric speeds (symmetric means upload and download speeds are roughly the same, as opposed to ADSL, where you’d be lucky to even get something like 1Mbps upload speed) that make even LTE look like dialup speeds. So, to properly verify that our speeds were indeed as good as they appeared, we tried some permutations of different test scenarios using our various devices. Between us we had 4 laptops and 3 smartphones, each with varying wireless network capabilities so they provided a broad enough range of devices with which to test the various access points.
The procedure for each of the access points was the same. We speed tested each access point on Ookla's speedtest.net using the Vanilla speed test servers, using both frequencies (5GHz and 2.4GHz) from different distances, testing 3 times at each distance, then averaging our findings. We first did a run of shot distance tests. Our first test was around 1m away from the access point, then 5m away, then 10m and finally we tested in the play room (around 15m away from the access point). During each test, only the person testing the network would be allowed onto the network. After those 4 test scenarios were complete, we all then simultaneously speed tested the network from inside the board room to simulate multiple users using the network (but as you'd expect, 3 interns don't quite do the same damage to the network as a whole company or home full of people would). After gathering those results, we then tested the devices over longer distance.
We were given 5 devices to test:
The Archer C2 TP-Link is built for a home user who likes reliability and quality. It isn't pretty, but it gets the job done. It has a removable antenna although I really would advise against removing them. The C2 has an added benefit - compared to these others - of being a WAN router. That means it can do things like host a DHCP service (so if you connect to the C2 you should be able to browse immediately, assuming the router is connected to the internet).
The Ubiquiti devices are all similar in design. They look like white discs and emit a low LED status light which caught my eye for longer than it should've. They're very lightweight and extremely portable, so you can slip them anywhere you see fit (perhaps under your bed for added security against the monsters?). Since they are just access points and nothing else, you'll still need a router or a router board that has a DCHP server running. They are enterprise products, so they tend to be a bit pricier, but that means that you can get some sort of performance guarantee under heavier loads (way more than 3 people). If you're like me, and you just want to know what to use in your house, ask yourself this: do we have 30 devices connected? The Ubiquiti products should cope better with many devices, since they've got settings for airtime fairness (which just means that one person can't hog the WiFi) and they support many more simultaneous connections. The Ubiquiti devices also require a controller software to manage them which although it's free, has it's own challenges for lay people. Using this software tool one can adjust various settings in order to optimise wireless connections for different locations and devices. It's not too difficult to use if you're even somewhat tech savvy, and there is plenty of support available.
As you can see, the differences between the devices are really slim in the short range tests. None of the devices really outperforms the others, so unless you're a really big family or multiple families or a small office of users, all of these devices are suitable for your needs.
Regarding the graph below, the lines at the bottom are the 2.4GHz frequency band, while the lines on top are 5GHz. What's immediately clear is the advantages of having a dual band router (meaning it can broadcast wireless signals at both frequencies). 5GHz is clearly faster, as long as you're sufficiently close to the access point. Surprisingly, the Lite actually out performs the Pro at greater distances, and this is more clear in the longer range comparison graph.
So if its speed you're after, you have to take into account which frequency band you're on. 5GHz provides faster data rates over shorter distances (when compared to 2.4GHz), but can't deliver quite as well as 2.4GHz at longer distances since higher frequencies have a shorter range.
Distance
You also generally don't want to be too far away from any wireless access point if you can help it. There are lots of devices constantly broadcasting wireless networks, and those wireless signals generate noise. It's like if you're trying to watch a movie but the other cinema goers have no regard for you. You want to hear what's happening onscreen, but the guy next to you decides that he needs to cool down and starts chewing on ice cubes, while the couple in front of you won't start professing their love to each other. Your laptop suffers from the same problem (albeit, with less anger) when it tries transmitting data over the busy 2.4GHz frequency. Devices such microwave ovens (not to be confused with alien technology), cellphones and even your neighbour's WiFi can contribute to the noise on the 2.4GHz frequency band. 5GHz is generally less crowded and if you have a newer computer or smartphone, chances are good that you'll be able to enjoy the perks of having a quiet network.
As an example the image below shows all the 'noise' in the Vanilla offices. We weren't able to eliminate these other access points while testing, so having the noise in the office allowed us to properly simulate a typical home or office environment in a bustling area.
Reliability
These access points were chosen to be tested because they're reliable and have shown to have the best longevity from the experience of the company over the past 5 years. They're also readily available in South Africa. If they weren't, then problems with the device would mean shipping it away and waiting weeks for it to be resolved (imagine having no internet access for weeks!). The Ubiquiti AC Pro is the most expensive of the lot, while the C2 is by the far the cheapest of them all. The Lite is much cheaper than the Pro and delivered comparable, arguably better, results, so it's better value for money.
The Long Range didn't really shine at long ranges, but it's still cheaper than the Pro. If you were on a tight budget, the C2 is by no means a bad product (it's actually really great) and does the job really well (especially since in South Africa, affordable home internet that'll push the limits of any of these products is just a pipe dream). You have to keep in mind that the UniFi access points are just dedicated access points, so they don't have any of the DHCP capabilities of the C2, and if you didn't have a router already installed, you'll have to fork out some more money for one (although, your ISP should've installed one when setting you up). The UniFi products do support more simultaneous clients, so if you're looking to host a huge party that offers free WiFi, you can't go wrong with either.
Aesthetics
The TP-Link is just not as stylish as the UniFi access points, and when compared it just looks like a drab muffin (as opposed to a delicious cupcake). That being said, it really does depend on the type of environment in which you are installing. If you don't plan to move the access points (or if you're in need of a really poor light source), I'd recommend installing the UniFi products since they're easily mountable on ceilings and they're extremely sleek with a pleasing status light when they're working. They also use Power Over Ethernet (POE) to get power, which just means that instead of needing a power source near the access point, you can supply it with power using an ethernet cable, which essentially neatens up the potential mess that comes with lots of cables and wires. It also means you aren't restricted by the locations of your power sources (unless you want to get an electrician to add some more power outlets), since the the ethernet cable that supplies power can be up to 100m long. All the UniFi models that we tested shared the same design, as seen below.
If you're just going to have the Unifi products on the floor or desktop, you may as well get the C2 since it manages to keep up with the other 2 access points, and you only really lose a metre or so in terms of range, and if you're that far out you would need another access point anyway (unless you're happy to deal with the degradation of signal). It's the best value for money product, and in fact, you could get 2 C2's and it'll still be cheaper than buying a Ubiqiti AC Pro. If you want the best you should get multiple Ubiqiti Lites
We've heard about the Google WiFi which is highly rated and apparently very simple to install at a cost of over R4000 for a set of 3 (excluding shipping and handling). We know that it's still not going to perform as well as three wired Ubiqiti AC lites.. This isn't yet available in South Africa yet.
At the end of the day, our results show that wifi access point performance drastically decreases as you move further away. If you have a small place to cover, get one TP-link. Buying the Ubiquiti APs will best provide that extra bang for your buck if you are going to have multiple wired APs to ensure a seamless speedy environment in a medium or large size home.
Authors: Oliver Makins, Ural Jonathan, Zakariyah Toyer, Kouame Kouassi