CakeQOS CakeQOS-Merlin

[brummygit]

the speed tesults wildly flucuate as well a buffer bloatscores worst i've had is C

I think I'd be trying it without the unlimited and set the Download Bandwidth to 75Mbps & Upload Bandwidth to 18Mbps - then you can try slowly increasing them until your bufferbloat gets affected.

You can also remove dual-srchost as the default value triple-isolate includes the dual-srchost functionality.

Use 90% of up/down, besteffort bridged-ptm ack-filter

Let me know.

I agree with those settings, besteffort is a default value which is a good starting point

 

[rgnldo]

the speed tesults wildly flucuate as well a buffer bloatscores worst i've had is C

Your connectivity infrastructure needs to be adapted. Go adding other options and comparing.

 

[TechTinkerer]

Use 90% of up/down, besteffort bridged-ptm ack-filter

Let me know.

Speedtest was done on wireless and did 90% result

 

[rgnldo]

Thanks @rgnldo - happy you are loving the simplicity. I do.

This is the reason for opting for CakeQOS. Dynamic functionality. It is as I mentioned, in this debate, I intend to organize a standard and establish a solution for high latency and network congestion.

 

[Wade Coxon]

Try

( 0Mbit | 0Mbit | besteffort | unlimited bridged-ptm ack-filter dual-srchost )

I get woeful bufferbloat results on my VDSL connection using those settings. Numbers in the several tens of ms.

Simply setting "bridged-ptm" and then entering valid rate limits works best on my connection.


However, in my tinkering tonight I have noticed that there is a problem with the rate settings.

I can no longer set my upload speed as "8.9" Mbit. The script says that 8.9 is not a valid number, however it has had no problem accepting that number in the past.
It also doesn't interpret 8900kbit as a valid entry either.

Can we get this reverted so that it will accept decimal places again please!

 

[brummygit]

Speedtest was done on wireless and did 90% result

That's not normal to see such low values when your connection speed was quoted in your screenshot as 80Mbps / 20Mbps. Is that a true sync speed or am I misreading it?

Are you testing with wireless or wired connection to your router? - make sure you are getting a really good connection to the router for tests (next to it if using wifi

 

[TechTinkerer]

That's not normal to see such low values when your connection speed was quoted in your screenshot as 80Mbps / 20Mbps. Is that a true sync speed or am I misreading it?

Are you testing with wireless or wired connection to your router? - make sure you are getting a really good connection to the router for tests (next to it if using wifi

Its wireless with phone doing tests...the speed is from router config screen in same room as router

i have set 75Mbit and 18Mbit as brummy suggested

 

[ttgapers]

Speedtest was done on wireless and did 90% result

Someone changed my theory on buffer bloat tests, but I NOW tend to use a wired connection into my router for best "results" that eliminates any issue(s) with your wireless network (interference etc.). Curious on your results from a wired box. At the end of the day you are testing buffer bloat performance of your router and not your wireless network.

 

[TechTinkerer]

Someone changed my theory on buffer bloat tests, but I NOW tend to use a wired connection into my router for best "results" that eliminates any issue(s) with your wireless network (interference etc.). Curious on your results from a wired box. At the end of the day you are testing buffer bloat performance of your router and not your wireless network.

What we need is a qos friendly version of spdmerlin that includes bufferbloat, so it can be directly run on the router cuttingout the wireless middleman to get an accurate idea of performance @Jack Yaz is it possible?

 

[brummygit]

Its wireless with phone doing tests...the speed is from router config screen in same room as router

i have set 75Mbit and 18Mbit as brummy suggested

I made an assumption that wherever you are is like me in the UK - the sync speed for the DSL modem dictates the connection speed. If you have a slower service which is throttled elsewhere by the ISP those are the speeds you need to set. Essentially you are matching your speeds to 90% of the slowest point.

 

[TechTinkerer]

I made an assumption that wherever you are is like me in the UK - the sync speed for the DSL modem dictates the connection speed. If you have a slower service which is throttled elsewhere by the ISP those are the speeds you need to set. Essentially you are matching your speeds to 90% of the slowest point.

It's EE in UK so i have to set 90% of those speeds quoted on my modem config?

 

[ttgapers]

What we need is a qos friendly version of spdmerlin that includes bufferbloat, so it can be directly run on the router cuttingout the wireless middleman to get an accurate idea of performance @Jack Yaz is it possible?

I was thinking the same, but I have a selfish desire to have it included in connmon over spdMerlin (as I don't use this one), as I care about line quality (connmon) over speed (spdMerlin), but would be good to get bufferbloat for both included. The links I provided earlier in the thread provide the links to various tests that might be best to integrate with.....

 

[TechTinkerer]

I was thinking the same, but I have a selfish desire to have it included in connmon over spdMerlin (as I don't use this one), as I care about line quality (connmon) over speed (spdMerlin), but would be good to get bufferbloat for both included. The links I provided earlier in the thread provide the links to various tests that might be best to integrate with.....

Tbh im not sure how to properly do speedtest would mean running a long ethernet to pc to get accurate test result i prefer doing it properly just a bit of a faff, have set 72Mbit & 18Mbit based on modem sync speed, not a fan of those amsterdam test servers on dslreport for accuracy

 

[TechTinkerer]

I was thinking the same, but I have a selfish desire to have it included in connmon over spdMerlin (as I don't use this one), as I care about line quality (connmon) over speed (spdMerlin), but would be good to get bufferbloat for both included. The links I provided earlier in the thread provide the links to various tests that might be best to integrate with.....

Hmmm well i havent used connman tbh im not fussed which app has it as long as testing works properly

 

[ttgapers]

Hmmm well i havent used connman tbh im not fussed which app has it as long as testing works properly

Np - but it's good to understand line quality, and how it impacts speed, latency etc. A faulty circuit with an occasional packet drop can have adverse effects even on the faster connection speed. Just a note.

 

[Andy1932]

The commands aren't working for me.
For example, I just tried to change the settings. This basically happens for all the commands.

ASUSWRT-Merlin RT-AC86U 384.18_0 Sun Jun 28 18:00:47 UTC 2020
admin@RT-AC86U:/tmp/home/root# cake-qos settings

#########################################################
##               _                                     ##
##              | |                                    ##
##    ___  __ _ | | __ ___          __ _   ___   ___   ##
##   / __|/ _` || |/ // _ \ ______ / _` | / _ \ / __|  ##
##  | (__ |(_| ||   <|  __/|______| (_| || (_) |\__ \  ##
##   \___|\__,_||_|\_\\___|        \__, | \___/ |___/  ##
##                                    | |              ##
##                                    |_|              ##
##                                                     ##
##      https://github.com/ttgapers/cakeqos-merlin     ##
##                        v1.0.1                       ##
##                                                     ##
#########################################################


#########################################################

CakeQOS-Merlin: Stopping
Broadcom Packet Flow Cache learning via BLOG enabled.
CakeQOS-Merlin: Starting - ( 100Mbit | 9Mbit | besteffort | docsis ack-filter dual-srchost )
Broadcom Packet Flow Cache learning via BLOG disabled.
Broadcom Packet Flow Cache flushing the flows

#########################################################

admin@RT-AC86U:/tmp/home/root#

 

[Wade Coxon]

The commands aren't working for me.
For example, I just tried to change the settings. This basically happens for all the commands.

Dunno about the other commands, but "settings" isn't a valid command any more. To get to settings, you just run "cake-qos".

 

[Andy1932]

Dunno about the other commands, but "settings" isn't a valid command any more. To get to settings, you just run "cake-qos".

Well, that certainly helps. Thank you!

 

[ttgapers]

Well, that certainly helps. Thank you!

"settings" is still valid once I looked through the script, but you must feed it your commands, meaning it doesn't launch the menu. to launch menu cake-qos!

No updates required.

Thanks again!

 

[Wade Coxon]

"settings" is still valid once I looked through the script, but you must feed it your commands, meaning it doesn't launch the menu. to launch menu cake-qos!

No updates required.

Thanks again!

Ahh good stuff then.

I checked by invoking cake-qos with a rubbish parameter, and "settings" wasn't listed as an option. Maybe the hint/error text needs a quick update too.