I wrote the following as part of a post to an discussion forum (I can't remember which one), and people told me that they liked this information and said that others would profit from it if I posted it on my blog, which I now decided to post.

The context: for homework of one MOOC, people were trying to download a large virtual machine image that was offered both via an official HTTP source, which was slow.

A person volunteered that already had downloaded the image volunteered to create a bittorrent to help fellow students.

Other students started saying that the bittorrent download was faster up to a certain point and, then, things slowed down a lot. Of course, they got faster to the point where lots of pieces were already well distributed and, then, only the everybody had already downloaded things and, then, only the seeder had the missing segments of the torrent.

Then I commented with:

Depending on your bittorrent client, you can add the original HTTP download as a "Web seed" and, this way, your client will download both from the peers and also from the web to fill in parts that are not yet available.

One such client is the Free (as in Freedom) qbittorrent, but I guess that others also have this option.

Soon after that, people started downloading things much faster.

blog comments powered by Disqus