RT-AC87R Adaptive QoS Settings

[Wes]

I just bought this router and I need help setting up the adaptive QoS settings. My connection is 20 Mbps download, and 2 Mbps upload. I need to know what numbers to use when setting up the QoS.

 

[Nullity]

You need to test and see what your real-world throughput maximums are. Then subtract 3-10% from the observed maximums and use the resulting values in your QoS config.

 

[Wes]

You need to test and see what your real-world throughput maximums are. Then subtract 3-10% from the observed maximums and use the resulting values in your QoS config.

Thanks for your help! Would I need to test the speeds on a wired connection? Or can I test the speeds on a wireless connection?

 

[bodean]

Wireless is fine


Sent from my iPhone 6 using Tapatalk

 

[Nullity]

With the wireless only being required to achieve ~20/2Mb, you should be fine.

Wired is obviously preferred though.

 

[bodean]

I usually take 20% off my upload speed, and leave the download speed set to 0 (unlimited) for adaptive QOS. Reason for 0 is, my comcast has a blast subscription. Seems to be 50mbps but sometimes will "blast" up to 75mbps. As its not always at 75, I feel like setting at 50 will cap it so I won't achieve maximum download speed. Have yet to find a better technique to counter this.

 

[bodean]

Any other input on this?

 

[bodean]

Guess not.

 

[Haggster]

nothing, except that it'd be interesting to know whether it's possible to enter a fraction of an mbit as a working value. as you can see, Wes has only 2mbit upload bandwidth. setting the limit to 1mbit would cut the upload bandwidth down to half of his potential speed, which would be unacceptable.

For people with more bandwidth it's easier. My approach is the following: I perform a lot of speedtests with the fastest available server which will always max out my connection. then I take the average upload value of my results and subtract 20% of that from my upload speed, always rounding down to the next lower mbit value.

for my download value I check the speetests to the same server but I check the one single all-time fastest value that I was able to get, subtract 1mbit from that and round to the next lowest mbit value.

in simple words that means I thoroughly test the maximum real life value that my line can get for downloads and only take 1.x mbit away from that to ensure that the limit is just below the maximum of what my line is capable of.

while at the same time cutting more from my upload, using the most common method for ensuring properly working QoS functionality.

The reason is that a saturated downstream has a way smaller effect on congestion than a saturated upstream. and the effect is that downloads close to the maximum attainable speed are possible but that my line will otherwise have perfectly working QoS.

I've even heard people saying that download rate limiting is pointless altogether and that only upload limiting has any real effects, but I cannot really believe that.