|
|
|
Written by Stew
|
|
Tuesday, 04 September 2007 |
|
Page 2 of 3
Traditional FTP downloading is often described using a tree model, with the user's PC acting as a leaf, the content distribution server the branch, the main host acting as the trunk. This works well until a branch is cut off, thus the end user fails to see any download action.
Torrent technology acts differently. Upon completion of the download we become servers. These servers are actually called 'seeders'.So assuming at least one person has the file, I can then download it from them. Aha. Sounds just like Kazza or Limewire or any other p2p. Wait. It gets better.
The file is cut into blocks which are then identified as metadata (or data inside the data). So for example, John has 100% of the file. He's a seeder. He gives me block 1 of the file. Jane doesn't have block 1 of the file so I give it to her. She has part 2, so John doesn't have to give me part 2 because I can get it from Jane. And so on. Through breaking the file into blocks we can then share the data in a much more intelligent way.
A FTP download is what I would call a 'single transmission'.
A Torrent download is what would call a 'collective transmission'.
Unlike FTP downloads where data flows in one direction, torrents mean simultaneous downloading and uploading. We each give each other the parts we need and therefore put less stress on the 'seeder'. Once we have 100% ourselves, we become a seeder and we give other 'leechers' (people with less than 100%) the parts they need, who can then pass it on to others who don't yet have that part. It's a much more rational way of moving data about. Especially with large files. This kind of architecture, by the way, is known as a 'Mesh Architecture'.
Here's a diagram that may explain it more clearly.
The main advantage of mesh architectures is this: the more users involved the better. With traditional FTP, more users means less bandwidth. With torrents, more users mean more bandwidth.
|
|
Supported by:
|
