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

[Audionut]

I read up on it but didn't understand why I would use 90% instead of 100%.

To make sure that the limiting factor to throughput is your router. This then allows the router to entirely decide what packets get priority, and always get sent or received with that priority.

And every speed test is different, should I go with the average value of my spdMerlin results, which has daily, weekly, monthly test results? Or go with a different speedtest provider.

I prefer to run manual tests using "continuous ping" under "network analysis" in "network tools". This will display a nice graph of results over time.

I then like to manually down/upload a large file from a known good server that consistently provides good speeds.

Start with no QOS. Run ping, down/upload file. If there is bufferbloat, the ping results will swing widely resulting in large "jitter" values.

Turn on QOS, adjust manual bandwidth settings until the ping results remain consistent and are largely unaffected by down/uploading.

I have observed a need to wait approx 5 mins after adjusting QOS settings, before running the next test.

Where are the WebUI Speed settings?

I'm pretty sure that means the manual bandwidth settings.

 

[jorgsmash]

I'm pretty sure that means the manual bandwidth settings.

Ok I'll have to give that a shot. So I'm modifying the percentages in each category between 85-95% to try and hone in on the best value. Do I do that for every category? Since you mentioned file transfer that's only one category but do I just use the same percentage for all of them? There should be a YouTube video on this lol.

Some other questions: The OP mentioned switching over to fq_codel, so I switched to that and wanted to make sure the additional settings that appeared were correct. I used the drop down menu to select Cable (DOCSIS) and it prefilled the box on the right to 18. The check box for 'ATM' says that a lot of DSL lines use that, so I left it unchecked does this look correct? If this is not set correctly what would be the outcome?

Lastly, the OP mentioned if you set a Gaming IP, it needs to be entered in CIDR notation. I have the IP of my console set in the gaming IP but it's not in CIDR notation. He shows he used a /30 CIDR for his setup but I don't understand why. This gives him a range of 4 ips. I don't understand why. Would I go with a /24 since my network is a /24 or would I stick with a /30 like the OP suggests?

Sorry for all the questions but I'm trying to get everything right and there's no way to know if it's right or wrong!

 

[jorgsmash]

I then like to manually down/upload a large file from a known good server that consistently provides good speeds.

I have a Plex Media server on my network, would streaming a video file from one device on my LAN to another be a good test method or does it need to be a remote server?

 

[zackattack784]

Ok I'll have to give that a shot. So I'm modifying the percentages in each category between 85-95% to try and hone in on the best value. Do I do that for every category? Since you mentioned file transfer that's only one category but do I just use the same percentage for all of them? There should be a YouTube video on this lol.

Some other questions: The OP mentioned switching over to fq_codel, so I switched to that and wanted to make sure the additional settings that appeared were correct. I used the drop down menu to select Cable (DOCSIS) and it prefilled the box on the right to 18. The check box for 'ATM' says that a lot of DSL lines use that, so I left it unchecked does this look correct? If this is not set correctly what would be the outcome?

View attachment 23876

Lastly, the OP mentioned if you set a Gaming IP, it needs to be entered in CIDR notation. I have the IP of my console set in the gaming IP but it's not in CIDR notation. He shows he used a /30 CIDR for his setup but I don't understand why. This gives him a range of 4 ips. I don't understand why. Would I go with a /24 since my network is a /24 or would I stick with a /30 like the OP suggests?

Sorry for all the questions but I'm trying to get everything right and there's no way to know if it's right or wrong!

As far as I know the automatic bandwidth setting doesn’t work. Set your up and down speeds manually.
For one IP use /32. If your game console is 192.168.50.5 input 192.168.50.5/32 for the gaming rule.

 

[jorgsmash]

As far as I know the automatic bandwidth setting doesn’t work. Set your up and down speeds manually.
For one IP use /32. If your game console is 192.168.50.5 input 192.168.50.5/32 for the gaming rule.

That brings me back to my previous question, do I go with the speed results from DSLreports which is around 100mbs avg from about 5 tests, or go with the spdMerlin test results which run like every hour or every day or something and they show 300+mbs every time. Thanks!

 

[jorgsmash]

As far as I know the automatic bandwidth setting doesn’t work. Set your up and down speeds manually.
For one IP use /32. If your game console is 192.168.50.5 input 192.168.50.5/32 for the gaming rule.

Just did an additional test to confirm. DSLreports shows 100mbs down, and spdMerlin on my router, ran right after the DSLreports, gets 400mbs down. 450 is what I pay for.

 

[dave14305]

Looking for volunteers on 384.18 alpha (AC models only for now) to test a beta version of FreshJR_QOS that makes it compatible with the new Asus categories. This is not backwards compatible with 384.17 or earlier at this time.

curl "https://raw.githubusercontent.com/dave14305/FreshJR_QOS/develop/FreshJR_QOS.sh" -o /jffs/scripts/FreshJR_QOS --create-dirs && curl "https://raw.githubusercontent.com/dave14305/FreshJR_QOS/develop/FreshJR_QoS_Stats.asp" -o /jffs/scripts/www_FreshJR_QoS_Stats.asp && sh /jffs/scripts/FreshJR_QOS -install

I would recommend everyone put Learn-From-Home as the bottom class in the QoS priority customization, since there's a potential conflict with the Streaming class as both contain category 4. Also the Game Downloads/Transferring class is on the bubble revert to Learn-From-Home since its priority is now adjustable in the Customize UI. There's no real benefit to Learn-From-Home, but it could become a good default category for Untracked traffic. As of now, I'm still moving Untracked traffic to Others for continuity sake. I can't see any evidence that Asus is managing untracked traffic at all, so I'm forcing it where it used to be in FreshJR.

Feedback appreciated. Once 384.18 goes final, I will look to create a new thread for this fork, assuming FreshJR doesn't reappear. But for now, this is the best place to get your attention.

 

[immi803]

@dave14305
On 384.17 and get this in system log

Jun  4 01:20:19 A.QoS: restart A.QoS because set_qos_conf / set_qos_on / setup rule fail

I just updated to 8.9 from freshjr 8.8

 

[Jack Yaz]

@dave14305
On 384.17 and get this in system log

Jun  4 01:20:19 A.QoS: restart A.QoS because set_qos_conf / set_qos_on / setup rule fail

I just updated to 8.9 from freshjr 8.8

This is not backwards compatible with 384.17 or earlier at this time.

 

[dave14305]

@dave14305
On 384.17 and get this in system log

Jun  4 01:20:19 A.QoS: restart A.QoS because set_qos_conf / set_qos_on / setup rule fail

I just updated to 8.9 from freshjr 8.8

What Jack said. But the setup rule fail is normal when QoS is initializing (even without FreshJR).

 

[zackattack784]

Just did an additional test to confirm. DSLreports shows 100mbs down, and spdMerlin on my router, ran right after the DSLreports, gets 400mbs down. 450 is what I pay for.

View attachment 23877 View attachment 23878

Something’s not right there. If I run a Speedtest across multiple sites (dslreports, speedtest.net, fast.com, etc.) I always end up with results in the same ballpark. I’d manually check a few different sites with QOS off and see what you get. I’d imagine it’s going to be in the 450 range.

 

[Sinner]

I haven't set the bandwidth manually yet. I read up on it but didn't understand why I would use 90% instead of 100%. Is it so any one device/traffic category doesn't steal all the bandwidth? And every speed test is different, should I go with the average value of my spdMerlin results, which has daily, weekly, monthly test results? Or go with a different speedtest provider.

From the OP:
"
1) Use manual bandwidth with limits set to 85-95% of your non-throttled speedtest results

To zone in on the ideal bandwidth within this range you need to be aware of these relations while testing with DSLreports Speedtest:

If you are getting poor bufferbloat grade, reduce WebUI speeds
If you are getting poor quality, increase WebUI speeds."

Where are the WebUI Speed settings? I checked in the Asus GUI settings for QoS, the SSH menu for QoS, and on the DSLreports Speedtest page settings.

My DSLreports speed test results are about 1/3rd of the speeds that spdMerlin gets. DSLreports gets about 100mbs while spdMerlin gets 300mbs+ every time.


View attachment 23867 View attachment 23868

Thanks!

the very reason your tests are all over the plac is because your NOT setting qos below your maximum speeds.. in order for QOS to actually do anything at all it NEEDS to be in full control of your connection and BE the bottleneck to prioritize the traffic.. if your ISP is the bottleneck then your might as well turn qos off cuz that's what you've got already.

 

[MarkyPancake]

Some other questions: The OP mentioned switching over to fq_codel, so I switched to that and wanted to make sure the additional settings that appeared were correct. I used the drop down menu to select Cable (DOCSIS) and it prefilled the box on the right to 18. The check box for 'ATM' says that a lot of DSL lines use that, so I left it unchecked does this look correct? If this is not set correctly what would be the outcome?

We use a DOCSIS broadband provider, but whenever I use fq_codel with WAN packet overhead set to 18 it doesn't perform as well as sfq. I keep reading that fq_codel is better to use, so I switched back to it recently, but set the overhead to 0, as that's a suggestion I've seen and so far so good. I always leave ATM unselected.

 

[QuikSilver]

Looking for volunteers on 384.18 alpha to test a beta version of FreshJR_QOS that makes it compatible with the new Asus categories. This is not backwards compatible with 384.17 or earlier at this time.

curl "https://raw.githubusercontent.com/dave14305/FreshJR_QOS/develop/FreshJR_QOS.sh" -o /jffs/scripts/FreshJR_QOS --create-dirs && curl "https://raw.githubusercontent.com/dave14305/FreshJR_QOS/develop/FreshJR_QoS_Stats.asp" -o /jffs/scripts/www_FreshJR_QoS_Stats.asp && sh /jffs/scripts/FreshJR_QOS -install

I would recommend everyone put Learn-From-Home as the bottom class in the QoS priority customization, since there's a potential conflict with the Streaming class as both contain category 4. Also the Game Downloads/Transferring class is on the bubble revert to Learn-From-Home since its priority is now adjustable in the Customize UI. There's no real benefit to Learn-From-Home, but it could become a good default category for Untracked traffic. As of now, I'm still moving Untracked traffic to Others for continuity sake. I can't see any evidence that Asus is managing untracked traffic at all, so I'm forcing it where it used to be in FreshJR.

Feedback appreciated. Once 384.18 goes final, I will look to create a new thread for this fork, assuming FreshJR doesn't reappear. But for now, this is the best place to get your attention.

So excited!

 

[Kingp1n]

Looking for volunteers on 384.18 alpha to test a beta version of FreshJR_QOS that makes it compatible with the new Asus categories. This is not backwards compatible with 384.17 or earlier at this time.

curl "https://raw.githubusercontent.com/dave14305/FreshJR_QOS/develop/FreshJR_QOS.sh" -o /jffs/scripts/FreshJR_QOS --create-dirs && curl "https://raw.githubusercontent.com/dave14305/FreshJR_QOS/develop/FreshJR_QoS_Stats.asp" -o /jffs/scripts/www_FreshJR_QoS_Stats.asp && sh /jffs/scripts/FreshJR_QOS -install

I would recommend everyone put Learn-From-Home as the bottom class in the QoS priority customization, since there's a potential conflict with the Streaming class as both contain category 4. Also the Game Downloads/Transferring class is on the bubble revert to Learn-From-Home since its priority is now adjustable in the Customize UI. There's no real benefit to Learn-From-Home, but it could become a good default category for Untracked traffic. As of now, I'm still moving Untracked traffic to Others for continuity sake. I can't see any evidence that Asus is managing untracked traffic at all, so I'm forcing it where it used to be in FreshJR.

Feedback appreciated. Once 384.18 goes final, I will look to create a new thread for this fork, assuming FreshJR doesn't reappear. But for now, this is the best place to get your attention.

@dave14305, awesome work and I'm testing now.

Does this setup look ok to you?

 

[Wade Coxon]

That brings me back to my previous question, do I go with the speed results from DSLreports which is around 100mbs avg from about 5 tests, or go with the spdMerlin test results which run like every hour or every day or something and they show 300+mbs every time. Thanks!

Might be a stupid question, but have you verified that your PC has a 1000Mbps connection to the router? If the router reports 300+Mbps, but your PC reports a value locked to 100Mbps, then that would suggest your PC is only connecting to the router at 100Mbps. Apologies if you have already addressed this, I looked back but didn't see that the question had been asked.

 

[dave14305]

@dave14305, awesome work and I'm testing now.

Does this setup look ok to you?

Yes, looks fine. The important thing for the script is to have Learn-From-Home be lower than both Video and Audio Streaming and Web Surfing.

 

[Kingp1n]

Yes, looks fine. The important thing for the script is to have Learn-From-Home be lower than both Video and Audio Streaming and Web Surfing.

I noticed under the legacy rule (Downloads_mark_up) under Upload, you have 853 port for both UDP & TCP...should 853 be removed under UDP since it only affects TCP?

 

[Sinner]

Yes, looks fine. The important thing for the script is to have Learn-From-Home be lower than both Video and Audio Streaming and Web Surfing.

whats with the learn from home talk? I thought merlin said it was in upcoming release?

 

[dave14305]

whats with the learn from home talk? I thought merlin said it was in upcoming release?

This is for 384.18 alpha.