This topic has been marked solved and closed to new posts due to inactivity. We hope you'll join the conversation by posting to an open topic or starting a new one.
Hi, just put the Beta onto my newly acquired XR500 and the feature that intrigues me the most is definitely the Connection Benchmark, as my last router (pretty old router) gave me a B/C on bufferbloat on DSLreports with crazy upload latency spikes and I wanted to improve that.
How much value should I even put into DSLreports bufferbloat testing or is it wildly inaccurate? (XR500 allowed me to get an A for bufferbloat consistently on DSLREPORTS)
Is the Connection Benchmark a more optimized way for our routers/network in particular to test bufferbloat?
Every test I run is A+ A+ A on connection benchmark, in order of Speeds, Ping, and Ping under Load, but I haven't done any extensive QoS congestion reduction slider testing aside from the 70/70 recommendation.. Just wondering if it's possible to get to an A+ for Ping under Load but I'm sure there's so many variables, my ISP being one of them. (Xfinity cable 1000 down 40 up plan).
Loving the 3.0 dumaOS 9999999x more than 2.0 already, thanks a ton 🙂 ping heatmap is sweet too
So yes in theory, as your base ping is quite low you should be able to make Ping Under Load lower for upload than what you have now. I wouldn't have thought those factors would affect it but it is certainly possible. Yes the XR1000 has basically already been through its own 3.0 beta testing with users and so it is in a more stable state than the XR500 3.0 is.
DSLReports runs directly to the device so there can be additional factors there, WiFi, cable, PC network adapter settings etc. Therefore Connection Benchmark is a better solution as it runs directly from the router and therefore bypasses those other variables giving a more accurate result. It depends on your base ping but as that is A+ then yes it should be possible, you will just need to experiment with the Congestion Control percentages to find what gives you that result. Really happy to hear you're liking it!
Ohhh I see, that makes sense. I didn't know that DSL uses the device and the Benchmark uses the actual router only, that's just a lot more accurate.
I notice too when I disable my Ublock origin adblocker my tests get way better.
I really appreciate all the work you Netduma guys put in on this, especially these forums, constantly replying to everyone.
Anyways, is bandwidth allocation something I should be concerned about adjusting when trying to perfect bufferbloat or can I ignore it for my situation? 1000 down 40 up plan as I mentioned before, with give or take 8-9 devices on my network. (All Wi-Fi smartphones and my hardwired PC, that's it)
It also BLOWS MY MIND that speedtest.net gives me my usual normal speeds but on the DumaOS software I get my QoS'd speeds, is that the goodput feature in action? Just crazy technology, my last router didn't even have QoS
Thank you very much! Best to whitelist the router interface on any Adblockers you use as it can interfere with how the interface performs. You can ignore that for bufferbloat purposes. Sorry could you clarify the last point? Connection Benchmark shouldn't show the speeds you've set for QoS - it's actually being improved at the moment as it's not reporting as high as it should.
Oh yeah, when I setup my router on the DUMAOS 3.0 firmware and factory reset it & take the speed test it always gives me around 400 down and 40 up, when my true internet speed is more around 900+ down and 40 up, I did try setting the congestion control to 'Never' with the sliders at 100%, but that didn't change anything. My connection benchmarks are consistently at the 400-440 download speed range, when my real download speed is 900+ on speedtest.net.
Hey Fraser, thanks for the reply. I posted an essay about an hour ago on this thread regarding upload bufferbloat and it got deleted somehow unfortunately.
Time to retype it on my phone 🙂
I did some research on different threads mostly on the netduma website regarding the Speed test portion of the Connection Benchmark and yeah, I'm not concerned about it, super minor bug imo, Ping under Load is the real juice I'm interested in.
https://i.imgur.com/zLy6H5Q.png <--- here's what I get pretty much everytime for Ping under Load, my Download and Idle are perfect for me, as they are always within 1-2 ms of each other, but that upload spikes to 100+ then hovers around 45-50, which still gives me an A, but I've tried everything to fix that initial spike.
I saw on another thread you or someone else mentioned using command prompt and typing in 'Ping 8.8.8.8 -t' to someone with a similar issue as mine and sure enough, it brought back an average of 10ms with a high of 29ms.
I think I followed what someone named 'TrayDay' (might be his name) mentioned which was disabling share excess upload and using the 'Set Bandwidth Speed' as a replacement for the Congestion Sliders, and just always leaving them at 100%, and while it helped a little, I can't say it was conclusive enough.
Any other recommendations to optimize that upload or is it an Xfinity ISP issue you think? THANKS again!
That initial spike I believe is an anomaly as it seems to just happen on the XR500 version of this feature. Have you used Always for Congestion Control and tested percentages, every 10% for example?
Yes and shockingly my most consistent slider position is 90-95% for upload, my very best is 40 MS average for Ping under Load. Nothing fixes the spike though.
One thing I forgot to mention was I disabled QoS and it got a lot worse, that spike you see at 100+ ms was what my upload CONSISTENTLY was when QoS was disabled, so the QoS is making more than a 50ms improvement to my upload bufferbloat, which is huge.
Do you think my upload bufferbloat being so poor regardless of where I test it (DSLreports, connection benchmark, etc.) could be because Xfinity cable doesn't support upstream docsis 3.1 yet?
Like for example I have an Arris surfboard sb8200 and the downstream bandwidth light on the modem is blue, indicating docsis 3.1 is active. But the upstream bandwidth light is green, indicating 3.0. Further research led me to find out that xfinity doesn't support 3.1 yet for upload, does that play any role?
Also, I went into my modems settings and checked it's signal and power stuff, everything checks out just fine except the Upstream bonded channel power is 41/42 when Arris recommends 45-51???? But then I read about having lower power is better? Could this be my issue or nah? Keep in mind I'm a huge noob at all this.
Is the XR1000 at the same stage as the XR500 in terms of its dumaOS 3.0 progression or would you say 3.0 is a little more stable on the XR1000? I might make the upgrade, not because I'm displeased with the XR500 though.
Seriously thanks again for reading these long messages, I think I'm out of questions.
So yes in theory, as your base ping is quite low you should be able to make Ping Under Load lower for upload than what you have now. I wouldn't have thought those factors would affect it but it is certainly possible. Yes the XR1000 has basically already been through its own 3.0 beta testing with users and so it is in a more stable state than the XR500 3.0 is.
Hey @Netduma-Fraser , sorry to revive this thread when you gave me the answers to my initial questions already, but I just got the XR1000 and the current firmware version I'm on is incredibly slow, any chance you could DM me a different 3.0 firmware for the XR1000? Thank you! When I say slow I mean like switching between QoS, Connection Benchmark, Geo-Filter, etc. the loading time is crazy long. Thanks again.
Awesome, thanks a ton. First time putting the DumaOS in action yesterday, never got into a CoD lobby above 30 ping using Ping Assist, wait times were less than 20 seconds. Incredible. Beyond satisfied
