[Release] FreshJR Adaptive QOS (Improvements / Custom Rules / and Inner workings)

[CriticJay]

I've installed this script a week ago on my 86U and it works exactly what a QoS system should be in my mind, so thanks for all the work you put in this project.

About your recommended priority list, right now my top priority is Gaming, so I put it above VoIP, but I have a doubt about Web Surfing and Video Streaming, is there any reason why Web Surfing should be above Video Streaming?

I believe video streaming doesn't need to be that high, because all video streaming services (Netflix, Youtube, etc) make use of buffering, so they can handle interruptions in the internet. Obviously VoIP/FaceTime + Gaming cannot handle interruptions.

If you put web surfing below video streaming then you might find yourself unnecessarily waiting longer for news web site / Facebook / Reddit to load, while your Netflix stream runs along fine. But your Netflix stream would run along fine REGARDLESS since it probably has a 10+ second buffer on it. So it's better to prioritize Web Surfing (Facebook, Reddit, Internet forums, Amazon.com) above Video Streaming.

 

[Ripper]

I believe video streaming doesn't need to be that high, because all video streaming services (Netflix, Youtube, etc) make use of buffering, so they can handle interruptions in the internet. Obviously VoIP/FaceTime + Gaming cannot handle interruptions.

If you put web surfing below video streaming then you might find yourself unnecessarily waiting longer for news web site / Facebook / Reddit to load, while your Netflix stream runs along fine. But your Netflix stream would run along fine REGARDLESS since it probably has a 10+ second buffer on it. So it's better to prioritize Web Surfing (Facebook, Reddit, Internet forums, Amazon.com) above Video Streaming.

That's what I thought, I'm a bit new on this so I just needed to confirm it, thanks.

 

[Vexira]

@FreshJR will ai mesh support that Merlin is implemeting have any effect on QoS and will this potentially delay the next release?

 

[RMerlin]

@FreshJR will ai mesh support that Merlin is implemeting have any effect on QoS and will this potentially delay the next release?

I don't see why it would have any effect. QoS is done at the router level, it's unrelated to nodes.

Routing-wise, AiMesh is just a fancy version of an Access Point.

 

[Swistheater]

FreshJR is stable right now on the test phase platform.

 

[Vexira]

I don't see why it would have any effect. QoS is done at the router level, it's unrelated to nodes.

Routing-wise, AiMesh is just a fancy version of an Access Point.

Ahh awesome, that's good to hear.

 

[brummygit]

I'm tearing my hair out over a weird problem I have at the moment. I'm wondering if anyone can help.

I've always used FreshJR's script on top of Merlin with excellent results. I run an RT-AC68U and have approx 55mbps download and 11mbps upload. I'm currently running on Merlin 384.12.

When I run a test with DSLReports.com speed test I see my upload speed at approx 75% of the configured amount. If I turn off QoS I see full speed but obviously higher buffer bloat etc. I can't work out what's causing this behaviour and have even fully uninstalled FreshJRs script and reinstalled.

 

[Vexira]

I'm tearing my hair out over a weird problem I have at the moment. I'm wondering if anyone can help.

I've always used FreshJR's script on top of Merlin with excellent results. I run an RT-AC68U and have approx 55mbps download and 11mbps upload. I'm currently running on Merlin 384.12.

When I run a test with DSLReports.com speed test I see my upload speed at approx 75% of the configured amount. If I turn off QoS I see full speed but obviously higher buffer bloat etc. I can't work out what's causing this behaviour and have even fully uninstalled FreshJRs script and reinstalled.

Did you set the bandwidth to 95% and did you set a per packet over head value?

 

[brummygit]

Did you set the bandwidth to 95% and did you set a per packet over head value?

Yes, everything as normally configured for years. It's very strange as it only seems to affect DSLReports.com and only when QoS is switched on.

 

[Vexira]

Yes, everything as normally configured for years. It's very strange as it only seems to affect DSLReports.com and only when QoS is switched on.

Did you test speed test.net

 

[CriticJay]

Yes, everything as normally configured for years. It's very strange as it only seems to affect DSLReports.com and only when QoS is switched on.

DSLReports speed tests are completely BORKED, in my opinion

Use FAST.COM and check "loaded latency" figure (milliseconds).
It should be no more than 3x-4x unloaded latency, if QoS is working correctly.

i.e. if unloaded latency is 15ms your loaded latency should be no more than 60ms (some lucky people may only get 2x unloaded latency!)

if QoS is disabled or NOT working correctly, then loaded latency would be something ridiculous, like 100ms, 300ms, 400ms, etc.

 

[Starrbuck]

Thanks for this explanation of fast.com. Attached is my speed without QoS for my 500/500 connection. I'm hesitant to use it if it's not really needed.

 

[QuikSilver]

Thanks for this explanation of fast.com. Attached is my speed without QoS for my 500/500 connection. I'm hesitant to use it if it's not really needed.

Speeds like that you shoudn't really need QOS.

 

[Swistheater]

I'm tearing my hair out over a weird problem I have at the moment. I'm wondering if anyone can help.

I've always used FreshJR's script on top of Merlin with excellent results. I run an RT-AC68U and have approx 55mbps download and 11mbps upload. I'm currently running on Merlin 384.12.

When I run a test with DSLReports.com speed test I see my upload speed at approx 75% of the configured amount. If I turn off QoS I see full speed but obviously higher buffer bloat etc. I can't work out what's causing this behaviour and have even fully uninstalled FreshJRs script and reinstalled.

when using dslreports, the issue sometimes is how their test is defined and not necessarily how much of your connection is being used. you can customize your test on Dslreports to use more push/pull which will result in a better connection test, sometimes their predefined test are buggy and you have to do that because their test are just generic for single connection types, they might not match your connection type 100% and also their server test sites might not work well for your location as well.

try fast.com if your latency is low then you should be good.

 

[fearz]

@FreshJR - Did you fix the problem that you checked with me? remember?

Quoting your private message:

@fearz​

could be pre-existing connection bug.

We will continue debugging after next script release.

If anything,

-turn off all devices
-turn off router
-turn on router
-wait 5 min
-turn on all devices

this should rule out pre-existing connection bug/glitch, (this bug is worked around next relase.)

--

there was whitelisted traffic prior to restart,
​

All the above wasn't working.

 

[brummygit]

when using dslreports, the issue sometimes is how their test is defined and not necessarily how much of your connection is being used. you can customize your test on Dslreports to use more push/pull which will result in a better connection test, sometimes their predefined test are buggy and you have to do that because their test are just generic for single connection types, they might not match your connection type 100% and also their server test sites might not work well for your location as well.

try fast.com if your latency is low then you should be good.

Thanks to those that replied regarding my challenges. I'm still puzzled as some speed test sites perform normally and some give reduced upload results. It's not killing me at the moment so I will patiently await the next version of the script which I know will have bug fixes, and then retest.

 

[Vexira]

I can't wait till the next release is available I'm keeping up my excitement for it, it might help me troubleshoot too see it's if it's my copper Line dying or something else that is the issue.

 

[Kingp1n]

I can't wait till the next release is available I'm keeping up my excitement for it, it might help me troubleshoot too see it's if it's my copper Line dying or something else that is the issue.

@Vexira, I haven't heard anything about a new release....do you think it will be soon?

 

[Vexira]

@Vexira, I haven't heard anything about a new release....do you think it will be soon?

I guess as much probably some time next month, at the latest.

 

[galbert]

Anyone that plays PS4 or CoD using this QOS? It looks like the same QOS as Advanced Tomato.