What's new

FlexQoS FlexQoS 1.2.4 - Flexible QoS Enhancement Script for Adaptive QoS

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

@dave14305
Thanks for the update.

I do have a dumb question.

I normally leave the AppDB at default settings. I've never actually had created a rule before until someone noticed all STUN traffic being identify as web.

When I was there creating the rule for STUN rule it automatically populated the mark & class which is pretty cool.

So, after trying to search for other rules, I noticed if I type in XBOX, it creates the appdb rule (mark: 08003E /class: Gaming). If I already setup a gaming rule under the iptables, do I need to create the appdb redirection rules as well? Is there any benfefit for creating these appdb rules?

I noticed there is also one for the Sony Playtation & NitendonSwitch plus a few others but I didn't create the appbd rule (currently limited to 32. Just curious if I need to create additional rules to improve my gaming or just leave as default?

Much appreciated.
 
@dave14305
Thanks for the update.

I do have a dumb question.

I normally leave the AppDB at default settings. I've never actually had created a rule before until someone noticed all STUN traffic being identify as web.

When I was there creating the rule for STUN rule it automatically populated the mark & class which is pretty cool.

So, after trying to search for other rules, I noticed if I type in XBOX, it creates the appdb rule (mark: 08003E /class: Gaming). If I already setup a gaming rule under the iptables, do I need to create the appdb redirection rules as well? Is there any benfefit for creating these appdb rules?

I noticed there is also one for the Sony Playtation & NitendonSwitch plus a few others but I didn't create the appbd rule (currently limited to 32. Just curious if I need to create additional rules to improve my gaming or just leave as default?

Much appreciated.
AppDB rules are useful if you disagree with the Class that the Trend Micro engine assigns to the application. If XBOX and Playstation and NintendoSwitch are already classified as Gaming, why would you change it?

An iptables rule will not be affected by any AppDB rules. So no need for both.
 
AppDB rules are useful if you disagree with the Class that the Trend Micro engine assigns to the application. If XBOX and Playstation and NintendoSwitch are already classified as Gaming, why would you change it?

An iptables rule will not be affected by any AppDB rules. So no need for both.
Got it ...that makes sense.

When I looked at STUN it was classifying as Net Control. In the meantime, I selected STUN to fall under W-f-H. I appreciate the information.
 
AppDB rules are useful if you disagree with the Class that the Trend Micro engine assigns to the application. If XBOX and Playstation and NintendoSwitch are already classified as Gaming, why would you change it?

An iptables rule will not be affected by any AppDB rules. So no need for both.
From the router's perspective... are iptable rules (typically) less CPU intensive than AppDB rules?
 
From the router's perspective... are iptable rules (typically) less CPU intensive than AppDB rules?
We'd need to ask the router. :rolleyes:

Since there's no direct way to measure that question, we can only speculate which requires less work. I would guess an AppDB rule requires less work. But adding extra filters in tc can create more work. Do less work. Be more happy.
 
So In short, The Application Name & it's Mark is predetermined via TrendMicro/Asus BUT we can use Custom AppDB rules to redirect them into a different CLASS if desired?
Wow... I think some of this is finally sinking in ;-)
 
Last edited:
@dave14305 Is it possible for future releases, in tracked connections, to be able to choose an application and that filter to still be applied when coming back to the flexqos tab?
It's really annoying, though probably of little importance, having the data refresh disabled due to having too many tracked connections. Having to go down and select the filter you wish to apply.
I know it it's not the end of the world, but it's just a suggestion for the future.
I can appreciate and relate to this request @dave14305 . I have an Rpi on my network that is part of ntp.pool.org. As you can imagine, I have lots of port 123 connections.
Being able to save the filter would be nice. As @ReDeiRe said, not the end of the world but a good suggestion.
 
Hey Dave, I noticed you mentioned previously about "you all ditching me for CAKE" & I've been doing my best to keep current with all the recent merlin alpha development.
It's my understanding that CAKE is not compatible with the Asus Hardware Acceleration.
Hence for ISP speeds >300Mbps it will obviously prevent clients from ever approaching the super fast speeds which approach Gbps.
So for non-symmetrical high-speed ISP that is +300Mbps isn't FlexQos still the most Viable option?
In other words... we are STUCK with you ;-)
 
Hey Dave, I noticed you mentioned previously about "you all ditching me for CAKE" & I've been doing my best to keep current with all the recent merlin alpha development.
It's my understanding that CAKE is not compatible with the Asus Hardware Acceleration.
Hence for ISP speeds >300Mbps it will obviously prevent clients from ever approaching the super fast speeds which approach Gbps.
So for non-symmetrical high-speed ISP that is +300Mbps isn't FlexQos still the most Viable option?
In other words... we are STUCK with you ;-)
Yes, but the people who usually need QoS the most are the ones with lower speed connections. They have choices now.
 
Yesterday I was testing Flexqos in a RT-AX88u v1.1. I have fiber and the supplier provides up to 1Gb symmetrical. I am receiving around 800/900 depending of the day. I wanted to improve the bufferbloat, because according to the tests many times I am getting score B (and just a few times ... score A).

Speed test. Laptop LAN connected to the RT-AX88u.

https://www.speedtest.net/my-result/d/288512049

The Bufferbloat test with the Laptop connected to the RT-AX88u.

http://www.dslreports.com/speedtest/67626069

The supplier (Movistar) uses pppoe with 3 VLAN (VLAN for internet, VLAN for TV and VLAN for VOIP). And the modem is not in full bridge mode, because it is not possible in the modem supplied. It is just bridged for the internet VLAN.

I tried to setup Flexqos with Adaptive Qos and manual setting 800DL and 800 UL. Maybe a better choice could be Automatic instead Manual, I don’t know.

I also chose fq_codel and the “wan packet overhead" setting as ethernet vlan.

After applying all changes ... I made a speed test directly using the router app and also using spdMerlin ... and the results weren’t so good. The DL was around 350 and the UL around 450.

Are these decrease of speeds normal?

Even after disabling Qos ... the speeds were not recovered. I had to reset the whole Aimesh to recover the gigabit speeds under the router speedtest app and in the spdMerlin script.

Is there any setup to improve the Bufferbloat using Flexqos without the huge decrease of the DL and UL speeds?
 
Last edited:
Yes, but the people who usually need QoS the most are the ones with lower speed connections. They have choices now.
Very true. And for once, I'm finding myself very fortunate.
With my current ISP package of 300/15 & me "only having" a RT-AC68U (which doesn't seem to be CAKE Capable).
I don't really see a reason to upgrade my router.
Even if I upgrade my router hardware & try some CAKE It would max out around 300Mbps.
Unless of course I upgrade BOTH Hmmmmmmmmm.
But I think I'd need a serious cash-injection, first.
With FlexQos I can just be Happy... as things are.
Thanks again.
 
so again is the develop channel still the latest version or right now the stable version ??
 
After applying all changes ... I made a speed test directly using the router app and also using spdMerlin ... and the results weren’t so good. The DL was around 350 and the UL around 450.

Are these decrease of speeds normal?
Yes, this is normal. The router is a bad place to run a speedtest, especially at your high speeds. Use a PC to run a speedtest. The router CPU is inadequate to get reliable results.
 
I think I remember that some people had an issues when they were not getting their full connection speed with a wired connection thru a PC. I have 400/20 using flex QoS and I currently only get 1/2 the speed. When I turn off QOS I then get my full connection speed but bad bufferbloat. Is this the issue with TrendMicro. I do not have any guest network or aiprotection on either.
I have an ax 86u using the latest Merlin firmware (stable). Any thoughts. Thanks in advance for your help.
 
Last edited:
Can someone explain me what is this "QUIC"?

Also... Is it in the correct category?

Thanks!
 

Attachments

  • Screenshot_20210310-005235282_1.jpg
    Screenshot_20210310-005235282_1.jpg
    84.6 KB · Views: 105
@dave14305 , don't worry I'm still here, I tried cake wasn't too sure if it was working, the software is to Alpha even for me someone who jumps on Alpha firmware always, so I'll be sticking to you like glue for a long while yet.
 

Latest threads

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top