CakeQOS CakeQOS-Merlin

[lilstone87]

If we can have a list to choose from for the 'Extra Options' (I don't know how else we would know about them, except by reading the thread otherwise), it would be the icing on the cake!

I agree with this. Especially for a noob like me. I had no clue of the options available. Reason I posted in this thread. As when you first setup everything, it ask you to input something, and honestly.. I didn't know if leaving it blank, and hitting enter, would work. Overall not a big issue at all. But something that would be nice in the future.

 

[ugandy]

On the initial release of this script, I did my testing here.

https://www.snbforums.com/threads/release-cakeqos-merlin.64800/#post-595227

With v1.0 of this script, nothing has changed with my 1Gbps symmetrical Fibre connection. Still getting up to 480Mbps down and 520Mbps upload with up/down inputs of 1000/1000 in the script and besteffort selected.

I also tested with autorate-ingress and found the speeds to plummet to 320Mbps down and about 380Mbps upload.

DSLreports showed A for all tests (without the script, A+ for all).

Fast.com showed 1ms, 5ms unloaded, loaded latencies without the script and 2ms, 18ms unloaded, loaded latencies with the script used. With autorate-ingress, the loaded latencies increased to over 24ms.

A final test I did with setting cake-qos download/upload speeds, both to 250Mbps, showed speeds of 240Mbps up/down and loaded and unloaded latencies of 2ms each.

I would agree with @ugandy that where speeds vary greatly, autorate-ingress will be the best choice overall (at least for cable ISPs), but the highest speeds possible is manually setting them for the current ISP conditions.

HTH.

Great job on how fast this has come together for all concerned!

If we can have a list to choose from for the 'Extra Options' (I don't know how else we would know about them, except by reading the thread otherwise), it would be the icing on the cake!

This seems ready for inclusion in amtm any moment now.

~500Mbps with cake would be bliss but i can't replicate it. could it be that my comcast connection is crappier than i believe? my best number was 360Mbps on ax88 when i set limits to match my contracted 600/15. BTW i'm having suricata off during this testing period. only: diversion/skynet ehich won't affect data throughput.
i wonder if i'd get baster speed if I change from diffserv4 to besteffort

 

[L&LD]

@ugandy, I believe it may be so. On customers' cable ISPs, I have worked for a couple of hours on networks to find out that they should be getting speeds like yours. In practice, the speeds possible during my time there were in the 220Mbps download range and at best, 8Mbps upload. I told them to switch ISP's. They said no choice. I told them I would move.

Glad you now have a method to alleviate most pinch-points though!

 

[lilstone87]

maybe i messed up my test, but adptiveqos/flexqos was unable to control my latency when UL was saturated. no packed loss, but latency was over 1000ms

Honestly.. That's why I mentioned Cake>FQ_Codel. As that is likely the downfall of using it. But for some of us, more options is nice. Also I think I remember with at least FreshJR's script, WAN<->LAN speeds were better for us with gigabit connections. I could be wrong on this.. and overall, I haven't used FlexQoS yet. So I can't comment on the type of speed's I can hit with it. I do get 930-940mbps down, by 35-37mbps up, from my provider.

Also I am a gamer, so the control/options FlexQoS offers, look good from afar, ultimately it's how everything works within Adaptive QoS with FlexQoS, that will matter most to me. I will test both, when I have time, and see what fits me best. I decided to run this first though.

 

[archiel]

I am still watching, trying to decide whether this will be of any practical benefit and have a few questions

1. All the speed tests mentioned (dslreports.com, Fast.com, nperf.com, etc) are designed to be run when the system is quiet, but presumably the purpose of Cake is to manage traffic when it is busy - how can I see if it has a real-world effect?
2. Given that traffic speeds will always be varying based on the local load on my ISP (Sky Broadband in the UK, theoretical max speeds set by Sky 80mbs/20mbs - max line speed (from router vdsl status 95mbs/24mbs) and results will also depend on distance to testing station (for dslreports, nearest node is in Amsterdam) is there any reason to prefer one particular speed-test over another?
3. Under no load, the dslreports figures are typically the slowest - expected given the distance- but even so the results are mostly A+,A,A+ and at worst A,B,A - as such is cake likely to add value?
4. Does anyone understand nperf.com? Not only does it report a fast link - much faster than any other test - it is giving results that are almost as fast as and occasionally even faster than the vdsl status figures for DS Actual Rate / NS Actual Rate rates - truly impressive??

 

[ugandy]

@ugandy, I believe it may be so. On customers' cable ISPs, I have worked for a couple of hours on networks to find out that they should be getting speeds like yours. In practice, the speeds possible during my time there were in the 220Mbps download range and at best, 8Mbps upload. I told them to switch ISP's. They said no choice. I told them I would move.

Glad you now have a method to alleviate most pinch-points though!

i actually can get ATT giga fiber but three things are holding me back: 1) i dislike att as much as i dislike comcast 2) need to drill hole to bring line inside the house 3) their 'modem/router' does not support bridge mode, i'd have to put it in dmz to keep using my ax88, as I heard last. i may be wrong. I'm hoping that by 2022 comcast will have symmetrical docsis 4.0
and for the most part if i got guaranteed 300/10 I'd be happy. don't know what i'd do with symmetrical 1Gbps. i already reached the end of the internet twice. maybe in a couple of years immersive VR will change this

 

[lilstone87]

my best number was 360Mbps on ax88 when i set limits to match my contracted 600/15.

Well I wasn't using my AX88U, but decided to run it today. I flashed Merlin's latest build on it. Factory reset it, besides initial setup with wifi. I haven't added anything except what was needed, to run this. I just changed my numbers to both 1000/1000, and ran two fast.com test, one on firefox, and one chrome. Download maxed out at 430-440mbps.

Which honestly if that's the max I can hit with the CPU maxing out. I rather not use limits that high. So I would rather use around 380mbps on the download side. I am curious if the same limits will happen with FlexQoS running.

 

[ugandy]

Well I wasn't using my AX88U, but decided to run it today. I flashed Merlin's latest build on it. Factory reset it, besides initial setup with wifi. I haven't added anything except what was needed, to run this. I just changed my numbers to both 1000/1000, and ran two fast.com test, one on firefox, and one chrome. Download maxed out at 430-440mbps.

Which honestly if that's the max I can hit with the CPU maxing out. I rather not use limits that high. So I would rather use around 380mbps on the download side. I am curious if the same limits will happen with FlexQoS running.

when i compared cakeqos to flexqos, i got 360 for cake and 450 for flex.
i though it was because flex only disabled runner, and cake disables both runner and FC.
in any case, you guys are getting way higher numbers for cake on fiber, vs me on comcast.

i think the max throughput of cake must also be affected by the line quality? it makes sense, does it?

 

[lilstone87]

in any case, you guys are getting way higher numbers for cake on fiber, vs me on comcast.

Sadly I am on docsis as well with my cable provider. What I found odd a minute ago. I set my download back to 380, and ran a test on fast.com, and on the download side. My loaded latency was a bit higher, vs setting my download limit to 1000, and maxing out the cpu.

 

[L&LD]

@ugandy I may be wrong, but there is no 'bridge mode' because AT&T will offer you an ONT/Ethernet connection, not a modem.

The symmetrical speeds plus the fact that you aren't sharing any bandwidth with your immediate neighbors is what will make your Comcast connection look like it's from 1996.

Is there a friend with AT&T Fibre you can test your router on?

And if by line quality you mean 'not shared' then I would agree with you that cake-qos, along with everything else, is affected too.

 

[ugandy]

I am still watching, trying to decide whether this will be of any practical benefit and have a few questions

1. All the speed tests mentioned (dslreports.com, Fast.com, nperf.com, etc) are designed to be run when the system is quiet, but presumably the purpose of Cake is to manage traffic when it is busy - how can I see if it has a real-world effect?

you can start multiple ongoing video streams, a large file download, a skype/zoom session going, while checking ping plotter for latency packet loss.

in my case my simple test is: does my citrix work connection dies/lags/slows to a crawl when my daughter is uploading massive video project file? no cake: yes; cake: no

 

[ugandy]

@ugandy I may be wrong, but there is no 'bridge mode' because AT&T will offer you an ONT/Ethernet connection, not a modem.

The symmetrical speeds plus the fact that you aren't sharing any bandwidth with your immediate neighbors is what will make your Comcast connection look like it's from 1996.

Is there a friend with AT&T Fibre you can test your router on?

And if by line quality you mean 'not shared' then I would agree with you that cake-qos, along with everything else, is affected too.

@L&LD 1996 , are we talking 56Kbps or less?
i miss that sound a dial up

ok. so it sounds that we established that cake+comcast tops at 330-430(?)Mbps and cake+fiber at ~500Mbps

 

[L&LD]

@L&LD 1996 , are we talking 56Kbps or less?

It depends on how many modems and telephone lines you had to use.

I quickly found out that download speeds were always sufficient (for their time). The upload speeds were always the issue when symmetrical ISP connections weren't possible.

 

[ugandy]

have we found out if cake still works with FC enabled and only runner disabled (like flexqos)?
what are the commands i can use to re enable FC and test?

 

[ugandy]

It depends on how many modems and telephone lines you had to use.

I quickly found out that download speeds were always sufficient (for their time). The upload speeds were always the issue when symmetrical ISP connections weren't possible.

well, surely the biggest issue was if a darned family member lifted up the phone.

 

[QuikSilver]

It depends on how many modems and telephone lines you had to use.

I quickly found out that download speeds were always sufficient (for their time). The upload speeds were always the issue when symmetrical ISP connections weren't possible.

Ah dial-up, those were the days....side story...my parents invested in an item called "catch a call" that was bought from Radioshack. Idea was to allow you to answer secondary line without being disconnected. Never worked perfectly though.

 

[ugandy]

have we found out if cake still works with FC enabled and only runner disabled (like flexqos)?
what are the commands i can use to re enable FC and test?

i did a simple test with "fc enable" vs "fc disable"
if FC is enabled, cake does not appear to work: pingplotter went red during spdMerlin test, stays green with FC off

 

[Smokey613]

have we found out if cake still works with FC enabled and only runner disabled (like flexqos)?
what are the commands i can use to re enable FC and test?

I am curious about this. I am running FlexQoS and both FC and Runner are enabled.

 

[ugandy]

from my test/understanding so far:
cake: both must be off (or else cake does not control latency, my test)
flex: runner must be off, FC can be on (as per flexos thread)

 

[Smokey613]

from my test/understanding:
cake: both must be off
flex: runner must be off, FC on

If that is true, how would I disable Runner? I would think if it was needed, FlexQoS would disable Runner on install.