I have a Virgin Media DOCSIS cable connection 1160Mbps/105Mbps and I am trying to reduce the impact of bufferbloat on the connection using SQM.
I have configured SQM using the OpenWrt guide and have managed to shave around 20ms off the loaded bufferbloat but I am still struggling to get A+ when using waveform.
Have I missed something in the settings!?
interface : eth1
Download speed (ingress)
880000
Upload speed (egress)
87000
Queueing discipline
cake
Queueing discipline
layer_cake.qos
Link layer
Ethernet with overhead: select for e.g. VDSL2.
Which link layer technology to account for
Per Packet Overhead (bytes)
18
Advanced Linklayer Options
Maximum packet size
2047
Rate table size
128
Minimal packet size, MPU (byte); needs to be > 0 for ethernet size tables
Linklayer adaptation mechanism
84
Which linklayer adaptation mechanism to use; for testing only
default (default)
I have tried reducing the down rate down to 84000 knits and the bufferbloat still remains at 5ms. Huge improvement but not A+.
What problem are you trying to solve - like illustrate with link to the report via https://www.waveform.com/tools/bufferbloat WIRED without SQM 20ms is relative, if base latency is 3s (eg you live on the Moon) it is crumbles.
Please connect to your OpenWrt device using ssh and copy the output of the following commands and post it here using the "Preformatted text </> " button (red circle; this works best in the 'Markdown' composer view in the blue oval):
Remember to redact passwords, VPN keys, MAC addresses and any public IP addresses you may have:
When load is applied to the connection the latency increases to around 40ms with a 10-15ms jitter increase.
Docsis is notorious for this.
My intention is to reduce this impact, which I have done by dropping nearly 200Mbps off the download and 20Mbps off the upload but even dropping the connection to 50Mbps, I still get 5-7ms bufferbloat.
