[AC86U] Wifi Speeds Lower Than Expect With 100Mbps DSL

[roguetr]

Things like QOS, Ai protection act as a bottle neck for speeds. Same goes for firewalls as well, to optimize a router it is useful to turn off all the features that aren’t needed as they act as bottle necks. In my experiences following those steps provided me with the best possible Wi-Fi speeds. But I do understand that changing these settings may be intimidating, that’s why I recommend saving a back up file before hand just in case things go wrong.

OP said they did a factory reset and most of those features are not default. If those features were a problem, they'd affect wired also, but really those overheads shouldn't create a bottleneck at these rates.

Once OP has a baseline for wireless LAN performance, and if there's definitely an issue outside of connected rates etc, which I'm not sure there is, then it's definitely worth looking at what _is_ configured.

Sent from my MI 5 using Tapatalk

 

[OzarkEdge]

Things like QOS, Ai protection act as a bottle neck for speeds. Same goes for firewalls as well, to optimize a router it is useful to turn off all the features that aren’t needed as they act as bottle necks. In my experiences following those steps provided me with the best possible Wi-Fi speeds. But I do understand that changing these settings may be intimidating, that’s why I recommend saving a back up file before hand just in case things go wrong.

There's a difference between leaving things OFF, i.e., the default settings, and playing around with the options. My point is that you can leave the defaults, configure the basic requirements as I noted, and achieve rated speeds, assuming the router/network and clients are healthy... without experimenting with all the options. Later you can layer on any additional features you want while monitoring the results as you go along.

OE

 

[Pabla]

There's a difference between leaving things OFF, i.e., the default settings, and playing around with the options. My point is that you can leave the defaults, configure the basic requirements as I noted, and achieve rated speeds, assuming the router/network and clients are healthy... without experimenting with all the options. Later you can layer on any additional features you want while monitoring the results as you go along.

OE

I do totally agree with that, was just giving some other suggestions as well.

 

[Krab19]

Thank you all for the replies!

I was able to convince one of my friends to let me borrow their macbook pro to run these test as you guys have noted. My PC was wired directly to LAN1 with the rate showing up as 1Gbps. I also changed the name of the WiFi to ensure the only device connected was the macbook pro. As many applications as possible were closed on both the PC and the macbook to keep their activity as low as possible.

Running iperf between my PC and the macbook pro, and keeping the various setting changes I have made:

Results with router set back to complete default:

Results with router set back to default, and QoS enabled:

Also, HW acceleration is on!

 

[roguetr]

From my understanding, that speed is typical for a MBP at those PHY rates. You can try running iperf with some additional switches to try and saturate the available bandwidth:

iperf3 -R -c IPADDRESS -w 1024k -P 3 -t 30 -i 5

And see if that bumps it up. I think you should see it somewhere around 600Mbps.

At those rates, how does the speed test fair now? The bottleneck should be the internet connection.

Sent from my MI 5 using Tapatalk