If you enjoy a lot of gaming it makes sense, as the stats suggest you are near capacity download
1 Like
These seem a tad high, could you post the result of a dslreports speedtest (see https://forum.openwrt.org/t/sqm-qos-recommended-settings-for-the-dslreports-speedtest-bufferbloat-testing/2803 for how to best do this for the forum) so we can get an estimate of reasonable shaper bandwidth settings, please?
I believe that on a DOCSIS network this needs to be 64 and not zero.
I would probably recommend cake and layer-cake.qos here as that will allow to configure per-internal-IP-fairness, so no single XBox should be able to spoil the fun for everybody else... But first the speedtests, please 
Here's the link
I haven't changed any settings yet
Did this test on PC wired
I can't believe that xbox's would pull that much. Amazing! I can't even imagine having 4 on at one time or will I be screwed.
Run that speed test a few times a day for a few days....
Jot down the time and result......
Seems like potential congestion issues on the upstream....
aka... docsis bus network ( your neighbours )
Off subject what's the best browser to use for this forum to be able to send me a notification ringtone or something letting me know that I have a response here?
Go it!
Where do you see that just curious? Someone is getting a flat tire. LOL
Question, did you disable SQM for this speedtest? If not please repeat with SQM disabled, as we want an estimate of the maximally achievable goodput to make an educated guess how to set the shaper limits.
nope... I'll do it now.
Sorry, I should have mentioned that, but the speedtest looks nice bufferbloat-wise 
Also really close to the theoretical limit:
Measured:
24.02/1.91
theoretical limits:
24991 * (1500-20-20) / 1522 = 23972.96 Kbps
1992 * (1500-20-20) / 1522 = 1910 Kbps
(I just realized you account for 22 bytes of overhead not 18, so I corrected the values)
No worries... almost done. Damm I can't spell or my fingers are just moving to fast
Survey says!!
1 Like
What time is it where you are at?
2:31pm est. USA
So from this measurment it seems the ISP docsis shaper is set to around:
26 * 1518 / (1500-20-20) = 27.03 Mbps
1.9 * 1518 / (1500-20-20) = 1.98 Mbps
Sounds believable for the downstream, as docsis ISPs tend to over-provision, but on the uplink I am not 100% sure whether that is the shaper of too much traffic from your neighbors
Anyway, based on these I would set the shapers to:
downlink 27 * 0.9 = 24300
uplink 1.9 * 0.99 = 1881
But those are close enough to the settings you used before so this will not change much
OK... Funny? Can it possibly be me considering that the cable guys came here and just put a splitter on one cable line to hook up the new xbox modem. So two modems one line.
So the next step is per-IP-fairness.
Here is the gist of the instructions:
To enable Per-Host Isolation Add the following to the “Advanced option strings” (in the Interfaces → SQM-QoS page; Queue Discipline tab, look for the Dangerous Configuration options):
For queueing disciplines handling incoming packets from the internet (internet-ingress): nat dual-dsthost ingress
For queueing disciplines handling outgoing packets to the internet (internet-egress): nat dual-srchost
taken from https://openwrt.org/docs/guide-user/network/traffic-shaping/sqm-details
In DOCSIS networks the coaxial cable operates as a "bus" that is quite a number of modems receivers are directly electrically connected to the same piece of copper. So sure you can compete with your own modem, but without others active as well that should not be a problem, I believe a docsis segment has up to 200Mbps? uplink capacity (certainly 100 is quite common), so two modems with low bandwidth plans should not congest the bus...
The interface name in qos is not the same name as my wan name does that mean anything? My wan is eth0.2 not eth0.
good question! well leave that for @moeller0 and the other gurus!
just don't put quirms in my cake, moving forward.... you've got a good toolset now.... some skills
the two key constraints are your link capacity / speed..... i.e. gaming is a little more synchronous.... so that upspeed is getting close to the limits... a youtube video or two and we start to squeeze it... and that's the second constraint.... competing traffic.... local we are addressing now.... somewhat... hopefully...... remote ( bus, isp ) we are keeping an eye on......
i'm not super happy with the milleseconds on the speed test..... but we shall see over a day or few whose been eating the cake
Good news also is that the fundamentals are sound.... at the router level..... you wouldn't have got much upgrading / changing......