Im using my ISP's (Centrylink's Actiontec C3000A) Modem/Router in transparent bridging mode and my speeds aren't being capped by QOS at all. My download ping under load is fine but upload ping is really high, probably because its not being capped?
I'm bit of a newb but maybe some of this information is helpful.
Speeds 140mbps Down 20mbps Up Connection Type: PPPoE WAN Type Selected: VDSL2 Bonding
Line Mode Selected: VDSL2-17A
ISP Protocol: Transperent Bridging VLAN ID Setting: Tagged-201
MTU Size: 1492
Congestion Control sliders are both set at 70% but aren't capping the speed. I also tried using the V2.3.2.104-DumaOS3.0-Beta firmware and have done 3 factory resets with no change. The tests were taken without any other devices on the network only PC connected via Ethernet.
Well I should have done this first... I am dumb and for not even trying a speed test website. https://imgur.com/a/XW8gRB0 Thanyou Fraser.
Quick question for checking bufferbloat on speed test websites like DSL reports it shows up for about a sec like it's going to test it then disapears and doesnt check or give me a score? Is this becasue my connection type or is there any other way to test bufferbloat? https://imgur.com/a/OY3gKjA
Glad it's all working! I'm not sure about that specifically but the Ping Under Load in Connection Benchmark is bufferbloat and it will respond to QoS so you can use them combined to find the best setting for the best result.
I have found this to be the case with some browsers. I usually have much better luck with dslreports/speedtest using Internet Explorer or the new Edge browser. Mine always seems to fail using Firefox or Chrome. Which is a little odd because the new Edge browser is Chromium based.
Well I should have done this first... I am dumb and for not even trying a speed test website. https://imgur.com/a/XW8gRB0 Thanyou Fraser.
Quick question for checking bufferbloat on speed test websites like DSL reports it shows up for about a sec like it's going to test it then disapears and doesnt check or give me a score? Is this becasue my connection type or is there any other way to test bufferbloat? https://imgur.com/a/OY3gKjA
@Netduma-Fraser wrote: Glad it's all working! I'm not sure about that specifically but the Ping Under Load in Connection Benchmark is bufferbloat and it will respond to QoS so you can use them combined to find the best setting for the best result.
That's weird because my upload still has bufferbloat even when QOS is set to on 😞
I tried even at 2mbps upload and it still had upload bufferbloat of around 60-90ms. How would I go about checking if it's an ISP issue if the downlaod is fine?
Hey, it seems like it’s not the solution for your problem, but I want to mention that there is another bug, that is causing the same problem. Just gonna write it down here so people wo have it, can find it unser this topic.
In my case the upload bufferbloat doesn’t work if I let the router do the „Vlan Tagging“ on the wan side. This is important for all my fellow Telekom customers who need vlan tag 7. I then let my managed switch do the vlan 7 (tagged on wan and untagged on lan) since then, the upload buffer bloat properly works for me 🙂
I read about somewhere if you have PPPoE or vlan tagging upload buffer didn’t work because of a bug, but that was a year old forum post. They talked about setting up another router just for the tagging and set the dmz to the xr500. I’m not totally sure how it all works because I’m a newb but I would try it out if I knew what numbers to type in.
A little off topic ish but seems like a good place to post this question. I have 49mbps down and 8mbps up, 12ms average ping, PPPOE, UK BT Openreach modem in bridge mode, VDSL connection. I can't get the upload bufferbloat anywhere near my average ping of 12ms without reducing the QOS slider to 33% on the DOWNLOAD.???? Can any network guru's explain this at all? I have QOS set to 33% (16.2mbps) on DL and 79% (6.3mbps) on UL. That gets me in the 20's for upload bufferbloat, changing the UL % doesn't make a lot of difference unless I set it to 95% in whcih case it goes up into the hundreds of ms. If I set DL any higher it jumps up to in the 40's and above 60% jumps up into the 70's/80's/90's. Not complaining would just like to know why I need to limit download bandwidth so aggressively to get the upload buffer bloat down to a reasonable level. Thanks
@Sky_ What do you mean by it doesn't work exactly with VLAN enabled?
@Jawwk doesn't look like you're doing anything wrong from what you've shown. Have you got done in multiples of 10% to see if it reduces?
@CoronalAurora Connection Benchmark does a test direct from the router and so is unphased by devices, this means it is a good benchmark of what kind of connection you're receiving from the ISP. In your scenario, 49Mbps isn't a lot and so it can be easily saturated to cause you lag - 4K stream on Netflix for example requires min ~25Mbps so half your available bandwidth. Throw in a few extra devices doing their thing and you're on your way to lag city. In order to prevent your network becoming saturated and causing you lag you're basically reserving bandwidth so that devices can't max it out. Every connection is different and so the percentage will also be different.
@CoronalAuroraConnection Benchmark does a test direct from the router and so is unphased by devices, this means it is a good benchmark of what kind of connection you're receiving from the ISP. In your scenario, 49Mbps isn't a lot and so it can be easily saturated to cause you lag - 4K stream on Netflix for example requires min ~25Mbps so half your available bandwidth. Throw in a few extra devices doing their thing and you're on your way to lag city. In order to prevent your network becoming saturated and causing you lag you're basically reserving bandwidth so that devices can't max it out. Every connection is different and so the percentage will also be different.
Hi Fraser, thanks for your reply but......you haven't really said anything I don't already know, what I was asking is why I have to seriously limit my DOWNLOAD bandwidth to improve the UPLOAD bufferbloat. Doesn't make sense to me.....I'm sure there's a good networking related reason, just wanted to know if anyone had some deeper knowledge...
@Jawwkdoesn't look like you're doing anything wrong from what you've shown. Have you got done in multiples of 10% to see if it reduces?
Yes, I have tried even lowering download limit and upload to 1% with no affect. I'm leaning more towards either a bug or ISP issue. Kinda sucks but I'll wait till next beta update, for now its a paper weight 😕
@CoronalAurora there is no reason I am aware of that requires you to throttle that much other than your speeds and how the ISP handles a saturated network by default.
@Jawwk so the ms literally doesn't move at all from Always 1% - Always 100%?
@Netduma-Fraser wrote: @CoronalAurorathere is no reason I am aware of that requires you to throttle that much other than your speeds and how the ISP handles a saturated network by default.
@Jawwkso the ms literally doesn't move at all from Always 1% - Always 100%?
@Netduma-Fraser thanks for your reply...I will contact my ISP and see if they can shed any light on it...just doesn't make sense to limit download for improvements in upload..I would understand if I had to take upload qos slider down but that doesn't affect the BB reading on upload for some reason, only by taking my download to 33% am I able to achieve closer to average ping on upload.
@Netduma-Fraser So my upload speed isn't capped when I set it. This is super odd and I'm having same issue at CoronalAurora where if I lower Download % my Upload bufferbloat is lower. Makes no sense but heres some screenshots.
@Jawwk I'm wondering if it is because you're making changes on the dashboard and there isn't even time in between changing a setting and running a test to properly apply. If you apply QoS on the actual page and then perform the test does it appear normal?
@Netduma-Fraser wrote: @JawwkI'm wondering if it is because you're making changes on the dashboard and there isn't even time in between changing a setting and running a test to properly apply. If you apply QoS on the actual page and then perform the test does it appear normal?
Yeah Im not really sure because I had it set for a whole hour and it looks like it uncapped by itself even though it says capped 3% and 29% you can see in screenshot, but even after 5 minutes it still hasn't capped my upload. Do you think this has something to do with VDSL2 Bonding (VDSL2-17A) that it what type dsl I have?
I doubt it, please try setting QoS on the actual QoS page not the dashboard then do a benchmark test on the benchmark page not the dashboard and see if it works then. After you made a QoS change refresh speedtest.net then test.
@Netduma-Fraser wrote: I doubt it, please try setting QoS on the actual QoS page not the dashboard then do a benchmark test on the benchmark page not the dashboard and see if it works then. After you made a QoS change refresh speedtest.net then test.
Okay well the factory reset fixed the upload not being capped, but it didn't fix the issue of having to cap download to 1% to get upload bufferbloat lower.
Yes I did and now I will just use the QOS page and Connection benchmark page separately in the future to not get that bug. Still didn't fix my upload bufferbloat being affected by my download cap but if the other guy had the same issue its probably due to our connection type: PPPoE VDSL.
