R2 speed issues

[fatal0Efx]

I've tweeted this, but figured the forum would be a better place.

I'm getting strange speed issues on the R2.  My ISP line is rated 1gbps (cable).  With PC connected direct to modem, I get 950+mbps consistently.  When going through the R2 it tends to fluctuate between 700-850mbps, sometimes hitting 860s.

I've disabled QOS with little to no effect.

Also, when I run the connection benchmark within the R2, it reports only 100-120mbps download speeds with a D rating.  This is very odd.  Why would internal benchmark be so low?

Attached are screenshots of the QOS being disabled as well as the benchmark vs. speedtest using PC app.

 

[Netduma Fraser]

The speeds you're getting through the R2 on Speedtest.net is probably the maximum you will be able to achieve through the router. 

Regarding benchmark, we are aware that it will give you lower results for those with very high speed connections and we're working to align that closer to what you're getting with other sites. Could you also see what speeds you get from here please? https://speed.measurementlab.net/#/

[fatal0Efx]

Why would Speedtest.net be expected to be lower through the R2, compared to MLab, which btw produces worse results.  Though that test flows through Miami whereas Speedtest is through Tampa (closer to me).

RE the internal benchmark, why would that be less?  I've seen posts where people have the exact opposite issue, where benchmark is really high on their high speed connection but websites are less.  Is that even possible based on what you're suggesting?

Also just to be clear, the speetest on the right in the graphic below is with a Windows app, not the Speedtest website.  Performing the test on the website yields similar results as MLab's site.

 

[Netduma Fraser]

58 minutes ago, fatal0Efx said:

Why would Speedtest.net be expected to be lower through the R2, compared to MLab, which btw produces worse results.  Though that test flows through Miami whereas Speedtest is through Tampa (closer to me).

RE the internal benchmark, why would that be less?  I've seen posts where people have the exact opposite issue, where benchmark is really high on their high speed connection but websites are less.  Is that even possible based on what you're suggesting?

Also just to be clear, the speetest on the right in the graphic below is with a Windows app, not the Speedtest website.  Performing the test on the website yields similar results as MLab's site.

No I mean, the speeds you got on speedtest.net which are high are probably the max speed you will get overall through the router. For the most part people are having the same issue as you, I think it is a combination of where mlabs (which is what we use) try to connect you to and benchmark also not reporting it accurately. We are aware of this and looking into it. 

[fatal0Efx]

Understood.  I never even considered the internal benchmark was actually going to a specific service provider.  Silly me.

So knowing the WAN uplink is likely only a 1gbps link itself so it'll never see anything above that, but I'm curious as to why the R2  would lose 100-200 mbps, even without QoS disabled.  Thoughts?

[Netduma Fraser]

In the lab and so a perfect environment we've tested at around 850-950 with QoS disabled, household speeds will vary as it is a real world environment. It's most likely the hardware in combination with the software. Once the Rapp store is properly up and running and you're able to remove features I suspect that might go up. We are always optimizing though so it may go up before then. What is your modem model? - Is it a pure modem?

[fatal0Efx]

I get much less than that. Again usually 700s-low 800s with R2, ~950 without. Seems def an R2 issue. 

Spectrum EU2251 (UBEE) DOCSIS 3.1 Modem in 3.1 state.  https://www.google.com/url?sa=t&source=web&rct=j&url=https://d15yx0mnc9teae.cloudfront.net/sites/default/files/SpectrumD31EMTA_New_Sheet.pdf&ved=2ahUKEwj2svDc_4fsAhVGuVkKHYo1AXEQFjAAegQIAhAB&usg=AOvVaw25lHv0SpWlYOW7mBSyJsAM

[Netduma Fraser]

Okay so your setup is fine then, will be what I said above in that case.