4 ms·

MOOC: Reducing Internet Latency: Why and How

3 points by osivertsson 3y ago

westurner 3y ago
Bufferbloat > Solutions and mitigations https://en.wikipedia.org/wiki/Bufferbloat#Solutions_and_mitigations https://en.wikipedia.org/wiki/Bufferbloat#Solutions_and_miti... AQM Active Queue Management: https://en.wikipedia.org/wiki/Active_queue_management https://en.wikipedia.org/wiki/Active_queue_management CoDel > Derived algorithms > CAKE: https://en.wikipedia.org/wiki/CoDel#Derived_algorithms https://en.wikipedia.org/wiki/CoDel#Derived_algorithms : > Common Applications Kept Enhanced (CAKE; sch_cake in Linux code) is a combined traffic shaper and AQM algorithm presented by the bufferbloat project in 2018. It builds on the experience of using fq_codel with the HTB (Hierarchy Token Bucket) traffic shaper. It improves over the Linux htb+fq_codel implementation by reducing hash collisions between flows, reducing CPU utilization in traffic shaping, and in a few other ways. [18] Re: the `dslreports_8dn` netperf test, bufferbloat, RRUL and the fluent GUI: https://news.ycombinator.com/item?id=26596395 https://news.ycombinator.com/item?id=26596395
- westurner 3y ago
  OpenWRT has a SQM "Smart Queue Management" package and LuCI web UI config that defaults to the CAKE scheduler. SQM is not installed by default. There is an x86 container with OpenWRT that can probably be used with load simulation and additional network interfaces on the container. IIRC I had to figure out how to start procd (~systemd) in the openwrt container myself. /? OpenWRT bufferbloat https://www.google.com/search?q=openwrt+bufferbloat https://www.google.com/search?q=openwrt+bufferbloat /? automatically optimize bufferbloat https://www.google.com/search?q=automatically%2Boptimize%2Bbufferbloat%2Bsite%3Agithub.com https://www.google.com/search?q=automatically%2Boptimize%2Bb... : sqm-autorate/sqm-autorate https://github.com/sqm-autorate/sqm-autorate https://github.com/sqm-autorate/sqm-autorate : > sqm-autorate is a program for OpenWRT that actively manages the CAKE Smart Queue Management (SQM) bandwidth settings through measurments of traffic load and latency. It is designed for variable bandwidth connections such as DOCIS/cable and LTE/wireless, and is not so useful for connections that have a stable, fixed bandwidth. speedtest-netperf.sh: https://github.com/openwrt/packages/blob/master/net/speedtest-netperf/files/speedtest-netperf.sh https://github.com/openwrt/packages/blob/master/net/speedtes... : opkg install speedtest-netperf speedtest-netperf.sh [-4 | -6] [-H netperf-server] [-t duration] [-p host-to-ping] [-n simultaneous-streams ] [-s | -c] You can run the Flent CLI in a Jupyter notebook with a heading for each test location. There are various ways to template Jupyter notebooks; e.g. `cp nbtemplate.ipynb wifitest_yyymmddhhss+z.ipynb`

3 points by osivertsson 3y ago

Bufferbloat > Solutions and mitigations https://en.wikipedia.org/wiki/Bufferbloat#Solutions_and_mitigations https://en.wikipedia.org/wiki/Bufferbloat#Solutions_and_miti... AQM Active Queue Management: https://en.wikipedia.org/wiki/Active_queue_management https://en.wikipedia.org/wiki/Active_queue_management CoDel > Derived algorithms > CAKE: https://en.wikipedia.org/wiki/CoDel#Derived_algorithms https://en.wikipedia.org/wiki/CoDel#Derived_algorithms : > Common Applications Kept Enhanced (CAKE; sch_cake in Linux code) is a combined traffic shaper and AQM algorithm presented by the bufferbloat project in 2018. It builds on the experience of using fq_codel with the HTB (Hierarchy Token Bucket) traffic shaper. It improves over the Linux htb+fq_codel implementation by reducing hash collisions between flows, reducing CPU utilization in traffic shaping, and in a few other ways. [18] Re: the `dslreports_8dn` netperf test, bufferbloat, RRUL and the fluent GUI: https://news.ycombinator.com/item?id=26596395 https://news.ycombinator.com/item?id=26596395

westurner 3y ago
OpenWRT has a SQM "Smart Queue Management" package and LuCI web UI config that defaults to the CAKE scheduler. SQM is not installed by default. There is an x86 container with OpenWRT that can probably be used with load simulation and additional network interfaces on the container. IIRC I had to figure out how to start procd (~systemd) in the openwrt container myself. /? OpenWRT bufferbloat https://www.google.com/search?q=openwrt+bufferbloat https://www.google.com/search?q=openwrt+bufferbloat /? automatically optimize bufferbloat https://www.google.com/search?q=automatically%2Boptimize%2Bbufferbloat%2Bsite%3Agithub.com https://www.google.com/search?q=automatically%2Boptimize%2Bb... : sqm-autorate/sqm-autorate https://github.com/sqm-autorate/sqm-autorate https://github.com/sqm-autorate/sqm-autorate : > sqm-autorate is a program for OpenWRT that actively manages the CAKE Smart Queue Management (SQM) bandwidth settings through measurments of traffic load and latency. It is designed for variable bandwidth connections such as DOCIS/cable and LTE/wireless, and is not so useful for connections that have a stable, fixed bandwidth. speedtest-netperf.sh: https://github.com/openwrt/packages/blob/master/net/speedtest-netperf/files/speedtest-netperf.sh https://github.com/openwrt/packages/blob/master/net/speedtes... : opkg install speedtest-netperf speedtest-netperf.sh [-4 | -6] [-H netperf-server] [-t duration] [-p host-to-ping] [-n simultaneous-streams ] [-s | -c] You can run the Flent CLI in a Jupyter notebook with a heading for each test location. There are various ways to template Jupyter notebooks; e.g. `cp nbtemplate.ipynb wifitest_yyymmddhhss+z.ipynb`