Leaderboard

Exactly! This is what most ISP's do. I've been saying this forever. I noticed this with P2P, (some) FTP and even trying to send a friend a large file via IM. I can actually prove it in DC++ hubs. Transferring file to another PC on my OWN network will be limited to a fraction of my "Gigabit" connection, unless the receiving PC has multiple upload slots, which will than have almost symmetrical speeds from the host pc (x) the amount of connections. Just about all residential ISP's shape network bandwidth and its infuriating. This is also done on cell phone data as well. Hence why they will "give you the option" or examples of their different "plan tiers". I have T-Mobile for example; If I do a speed test I can pull lets say 100/80 , but if I try and play a YouTube video @ 2160p all of a sudden I have buffering @ 332kbs.... Makes no sense right? Now if I drop the resolution to 1080p , it'll almost instantly download the video @ 18mbs. lol I believe they take the known ports or name servers from certain applications and limit them intentionally, whereas speed tests are just raw data. idk. I just work here.

Thank you for sharing this information. It is very helpful to understand the challenges of measuring upload speeds and the limitations of current speed test tools. It is also encouraging to know your working on a new version of TestMy.net that will address these challenges. I am looking forward to testing the new version and seeing how it performs.