Instead of just testing for the maximum throughput from a single laptop, we used six laptops, spaced around our New York test facility, to simulate the real-world activity of a busy home network. The cellar of our office building has the luxury of space, as well as a mix of Wi-Fi challenges: masonry walls and drywall construction, open spaces, glass windows, and metal-framed doors. We used a TP-Link Archer A7 (our budget router pick at the time of our testing) as the baseline for our tests; it’s an excellent router in a small space such as a townhome or apartment, but its signals are stretched a little thin in our test space.
Because these tests simulated real-world traffic, they did a better job of modeling everyday performance compared with a tool like iPerf, an artificial testing tool that moves data from one machine to another as fast as possible. We did similar testing for the latest version of our guide to Wi-Fi mesh networks.
Test device placement
Putting devices in the right places is key to any mesh network’s success; you should space them out in a way that gives all areas of your home Wi-Fi coverage. We started by placing the router in our office’s prep kitchen, in the center of our testing space, connecting it to our Internet connection via Ethernet. We put each extender about 40 feet away from the router, in our lounge on the other side of the office’s main kitchen, well within the “bubble” of the Wi-Fi signal from the Archer A7. Remember, despite their name, signal boosters can’t extend the signal too far beyond where the router’s original signal gives out.
During testing, the six laptops, our wired controller laptop, and an Apple iPhone running the extender app (if needed) were the only devices connected to the test network. We didn’t disable any of the surrounding Wi-Fi networks or wireless devices like Sonos speaker systems; these kept doing their usual noisy things, just as they probably do in your home. The neighbors and our corporate network also kept their Wi-Fi networks going, which left somewhere in the vicinity of a dozen network names visible at any given time.
We spaced the six client laptops so that they should naturally try to connect to the router or extender, whichever was closer, but we also grouped them close enough that some of the laptops could switch if necessary.
We updated the drivers on each extender, on the Archer A7 router, and on each client laptop (using Windows Update) before the testing sessions started. Upgrading your firmware to the most recent version after setting up your network devices is crucial for receiving both potential performance improvements and security patches.
Our six laptops ran the following tests:
- One laptop downloaded a very large file. We wanted to see an overall throughput of 100 Mbps or better, to simulate the experience of an impatient person waiting for a device to complete an update. This test is a big challenge for the rest of the network—if this laptop gets all of the available airtime, the other tests suffer.
- Two laptops each simulated a 4K video streaming session. They tried to download data at up to 30 Mbps, but we were satisfied if they could average 25 Mbps or better, which is what Netflix recommends for 4K. If either laptop can’t get at least 20 to 25 Mbps, that means a real video would be pausing and buffering.
- Three laptops simulated real-human Web browsing by loading a “Web page” once every 20 seconds. Each “Web page” consisted of 16 separate 128 KB files, all requested simultaneously, and we measured latency from the time the requests went out to the time all 16 requests were fulfilled. This is the most important test—it accurately represents the thing that frustrates real users most (slow and inconsistent Web browsing)—and it usually fails before any of the other tests do.
We ran all these tests at the same time for a full five minutes to simulate a realistic overly active time on a home network. Although your network probably isn’t always that congested, it is that busy often enough—and those are the times when you’re most likely to get annoyed, so that experience was what we were modeling in our tests.
These tests simultaneously evaluated range, throughput, and the network’s ability to multitask. We connected each extender to the Archer A7 router via Wi-Fi and then created a separate Wi-Fi network with a unique SSID. If the extender was mesh-compatible, we used a single network name (or SSID) to enable roaming between the router and extender. We didn’t touch most of the other settings—you should be able to connect to your Wi-Fi and have it work without constantly fiddling with things.
Since we were testing extenders for your existing router, rather than a complete replacement, we measured our results by improvement rather than raw numbers. Before testing any extenders, we ran baseline tests in exactly the same way using only our Archer A7 base router. Then, as we tested each extender, we subtracted the value of our baseline test. This approach allows us to directly show you how much each device improved—or degraded—our test network’s performance from what we started out with.