WiFi Upload/Download Speed Differences

[Bandito]

I’ve not seen this addressed online and am very curious as to what may be going on.

I have a 1 Gbps symmetric fiber connection to my home, but my WiFi upload speeds are always half of the download speed or a little less, i.e. 400 down/125 up. This has been true ever since I got the fiber connection. It doesn’t matter which WiFi system I use, whether its the provider’s system, Netgear Orbi, Linksys Velop, or ASUS XT8. They all perform more or less the same with the upload speed being significantly less than the download speed.

A hardwired connection always provides me with a symmetric upload and download speed, so I know that the overall connection is functioning as it should, it’s just the WiFi that shows this kind of imbalanced result.

So, my question is, is there something inherent in the WiFi specifications or implementations that causes this apparently inherent difference in speeds?

Thanks for any enlightenment that you can provide!

 

[bonerking]

Have you tested different wifi clients? Does it happen even directly next to the router?
At some distance from the router I get roughly this behavior on my intel wifi laptop, but our broadcom wifi iPhones do much better, sometimes the upload is even greater than the download, and they will easily hit 500/500 (line cap).
Close to the router the laptop reaches equivalent up and down speeds.

So, my question is, is there something inherent in the WiFi specifications or implementations that causes this apparently inherent difference in speeds?

The main things I can think of are
1. Router having greater TX output power / gain than the client.
2. Router supporting beamforming but client doesn't.

 

[Bandito]

Yes, I have checked with several Windows PCs, Macs, iPhones, and iPads. All do the same thing and remember that this has held true over the use of at least five different WiFi systems from different manufacturers.

Getting close to the router will increase the speeds that I see for both uploads and downloads, but the difference between uploads and downloads remains the same.

Also, I’m using the Speedtest app for checking speeds. I’ve tried several different apps and websites and Speedtest seems to be the most reliable and consistent test for me. I have also tried dozens of different servers to test against, but with the same general results.

Thanks for sharing your thoughts!

 

[cplay]

wifi is half duplex (ethernet is full).

most of your clients will be 866mbps link rate (2 spatial streams) so you will get roughly half that in throughput - which you are.

Focus on more important things, life is short, the difference between 400-500 means jack sh*t for normal use.

 

[Trip]

wifi is half duplex (ethernet is full). [...] most of your clients will be 866mbps link rate (2 spatial streams) so you will get roughly half that in throughput - which you are [...] the difference between 400-500 means jack sh*t for normal use.

The OP is not griping about that; he's specifically wondering why his internet-based wireless upload speeds are roughly 25% of download speeds. And wifi being half-duplex is a pretty broad explanation.

@Bandito - First off, you want to speed test your wireless link locally, making sure you have only a single wireless client connected to wifi, testing to a single wired host, directly connected to the AP if possible, running something Duckware speedtest. This will ensure your measurement is free from any throughput limitations placed on egresssing that wireless traffic further up-network and out to the internet, which can happen quite often.

If you're still noticing a mismatch when testing locally, your clients are most-likely selecting lower MCS encoding rates for Tx than Rx to the AP/station, which is in accordance with their much weaker radios, and/or extra RF interference in your environment causing that selection to occur, which would explain the behavior irrespective of AP product used (slight differences between products notwithstanding, of course).

For more clarity on this issue, see this Duckware article Appendixes E, F and G.

 

[Bandito]

@cplay - I’m not worried about a difference of 100 Mbps in my connection speeds, I am trying to learn why the upload speed is significantly slower than download speeds: usually half as fast or less. I do realize that WiFi is half duplex, but that doesn’t explain why the upload is so much slower than the download. And, I’m asking this out of curiosity and wanting to better educate myself.

@Trip - Thanks for better understanding my question. I haven’t done any local/LAN testing recently, but will give it another try to see what kind of speeds I’m getting internally on my network. Your reply that the clients’ radios are much weaker may explain a lot about the situation as well. I’ll read through the entire Duckware article when I have a few minutes.

Thanks to both of you for responding!

 

[Bandito]

I did the reading of the Duckware article that @Trip linked to in his response and picked up a few tidbits here and there.

I’m thinking that a lot of the difference in speeds is due to the clients having limited transmit power compared to the router itself. According to the Duckware article the clients typically max out at 250 mW whereas the routers are typically around 1000 mW, so that could explain the difference. I’ll have to do some more speed tests in close proximity to the router to see if this pans out or not. I also need to do some internal WLAN testing to see what’s occurring away from the Internet as well.