Starlink upload speed - TCP CUBIC vs BBR congestion control

[Sean]

Having recently moved from a deteriorating 4G service to Starlink, I tried some experimenting to see if I could get some extra upload speed.  From what I've found out in this presentation, Starlink has significant uplink packet loss as well as varying latency as the satellites fly past, i.e. lowest latency directly overhead and highest when going out of view towards the horizon.

 

As the BBR congestion control does not use packet loss as a congestion indicator and adapts well to varying latency such as cellular networks, this should work well with Starlink uploads.  The Ookla Speedtest makes multiple simultaneous uploads, so it generally reports much faster upload speeds than uploading a large file, however, with the Linear test on TestMy (i.e. Multithread disabled) the difference from my testing is huge.  So far I had no l luck finding any way of configuring Windows 10 or 11 to use BBR, however, this is possible with Linux with Kernel 4.9 and higher, such as much recent Linux distros.  While Windows 11 supports the BBR2 congestion control, it seemed to have little to no effect on upload speeds.

 

With a Linux root terminal:

modprobe tcp_bbr

 

Add the following lines to the end of /etc/sysctl.conf

net.ipv4.tcp_congestion_control = bbr
net.core.default_qdisc = fq

 

The following is a set of 5 upload tests across 5 different servers at varying distances from my location in Ireland, the final being Sydney at the other side of the globe.  I tried running a 50MB block in each, but after Germany, the CUBIC tests failed with a "Too Slow" message 😁, so used smaller blocks for those CUBIC tests.  Each upload test was run on the same Windows 10 PC (CUBIC congestion control) with Linux Mint running in a Hyper-V virtual machine configured for TCP BBR.

 

As the TCP BBR congestion control only affects uploading, for faster download speeds the remote server must use the BBR congestion control.  

[Jose Fernandes]

Hello.

 

Sean, Family and friends.

Some 4G providers with outdoor routers have better performance than others. The latency might be 50 ms.

Starlink is better than the 4G. 

5G FWA AirFiber are available in some countries with excellent performance.

In my area, we have 5G from a tower 500 meters away.  My smartphone is inexpensive (200€) made in 2021. The average upload download is 100/ 500Mbps.
In a few years from now I might have a samrtphone with Xring CPU with 3 nm from Xiaomi.

In AP (Acess Point) mode to the notebook the upload is 70 Mbps and the download is 240 Mbps. 

https://testmy.net/db/-52WFcdHv.x6hxRxL5P

 

We have Fiber Optic only since 2015 and also have ENYA CD.

 

Starlink is a good choice for the deserts in Australia, the poles, rural US, Africa, Brasil - remote areas.

Congratulations for your Starlink  where Halloween strarted.
 

Have a Beutiful Day

 

 

[CA3LE]

On 6/9/2025 at 2:13 PM, Sean said:

As the TCP BBR congestion control only affects uploading, for faster download speeds the remote server must use the BBR congestion control.  

 

Yes, I use BBR congestion control on all the test servers.

 

[root@dallas ~]# sysctl -a | grep tcp_congestion_control
net.ipv4.tcp_congestion_control = bbr

 

BBR was created in 2016 and I started testing with it immediately.  It quickly became the default here.  Before that TMN used cubic.

[Sean]

From further experimentation, it looks like I can take advantage of the Linux BBR congestion control in Windows via a Socks5 Proxy.  In Linux, I installed MicroSocks and ran it in a terminal for my test.  In Windows, I installed a trial version of Proxifier and configured it with the IP address from Linux.  A few TestMy tests on Windows 10 started giving me speeds just like on Linux:

 

pCloud which usually only gives me around 0.75MB/s (i.e. ~6Mbps) literally got a 10x improvement with Proxifier routing it through the Linux proxy server.  I was beginning to doubt the 8MB/s figure in the pCloud window until I brought up the Ethernet graph:

Here is a YouTube video where someone tried improving their upload speed by bonding 3 Starlink kits, obviously unaware of what upload speed they could have got with a single Starlink just by using Linux with the BBR congestion control . . .

 

[Frank225]

Hello Mr. Sean,

 

Congratulations for your hard work and eautiful work.
 

You deserve a cruise or make a long trip on a sailboat with one antenna and don’t worry about EMI and jammers.
 

And maybe buy a house in the deserts of Australia and learn to play the didgeridoo and buy a plastic one. There are nice pandas and other nice animals and plants.
 

Do you recommend your project to anyone with knowledge?
 

I prefer geostationary satellites at 36,000 km above Earth with laser communications in the near future for Africa and Asia.
 

I used to have:

• 4.20 meter parabolic dish, solid with high efficiency “53 dB @ 12 GHz for “geostationary satellites” for TV made by DH Satellite in the USA.” With a Drake receiver, low-threshold, low-noise LNB, and a good feedhorn, low-noise RG6 coax cable.
• Four high-gain Yagi antennas stacked vertical and horizontal (“H”) for 145 MHz communications over the horizon.

 

In my location, we have fiber optics, and my ISP gives 1000/400 Mbps and 1000/150 Mbps on 5G.
 

Very soon, 7-8 Gbps and 10 Gbps will be available in some countries.
 

Before, we could have more than one ADSL line, 3G, or 4G and make load balancing with inexpensive TP-Link load balancing.

ASUS routers with Merlin software and some Xiaomi can make load balancing. Xiaomi makes 3nm chips.
 

My last toy was an inexpensive 5G router.

 

enjoy

 

[Frank225]

For some reason I cannot put a like  💙 on Mr Sean for is lovely work in the wilderness.

 

[Jose Francisco]

Good sunday to everyone out there.

My upload is not the best in the west and we will relax until the next upgrade. 

In 2015 my Fiber Optic Internet Upload was only 3 (three) Mbps and 30 (thirty) Mbps Download.

ADSL, VDS, 3 and 4 G Users please be cool and have a better day tomorow ☮️