The further away from the server you are will affect the download speed from that server.
example: the client has 25mbps fibre link if the server is in SA ie 20ms away they can get full speed, however, if in London they only will get 3.6mbps
See: https://www.switch.ch/network/tools/tcp_throughput/
- complete the bottom section to determine the theoretical maximum Calculate Bandwidth-delay Product and TCP buffer size
Using web-based speed-test is generally not recommended for testing speeds over 100Mbps. The results are largely inaccurate. You can try using a speed test app for more accurate results.
https://www.speedtest.net/apps/windows
The best tool to confirm the speed of the line is to use iPerf. As we all know speedtest.net isn't the answer to everything we have no visibility on the endpoints as such you can't say what the true speed is or is not.
https://iperf.fr/iperf-download.php
Our servers have 10 ports open for TCP tests only ports 5200 to 5209
The server resets and restarts the iperf3 daemons every hour at 59 minutes into the hour, so don't be surprised if a test at the top of the hour fails or hangs, just cancel it and restart the test. The reset takes between 30 and 60 seconds to complete.
Server addresses :
Ports: 5200 to 5209 TCP only
Download test: iperf3.exe -c lon1.as37497.net -p 5204 -P 10 -t 20 -R
Upload test: iperf3.exe -c lon1.as37497.net -p 5204 -P 10 -t 20
Parameters :
-p Port can use any port between 5200 and 5209
-P #parallel tasks increase based on distance suggest 10 for local, 15 for London and 20 for further. This is due to tcp window size scaling.
-t time to run in seconds increase this based on distance, suggest 20 seconds for local and 30 to 40 seconds for international.
-R reverse ie receive data / Download