To do this, the traditional P2P model (where downloaders connect only to the file's host) has been modified so that the majority of your download is actually received from other downloaders, which eases congestion considerably.



The following terminology is useful to refer to when reading - 'The downloading process' below:

1. Client

Software that allows you to connect to a BitTorrent tracker and download a file from others on the network.

2. Tracker

An Internet server which manages the connections between the clients. No part of the actual download is ever transferred through the tracker however, nor does it hold a copy of the file.

3. .torrent file

Everything available for download using BitTorrent has a corresponding .torrent file. This is a small 'pointer file' (also known as a metadata file) that connects your client to the correct tracker, and links you up with other people sharing the same download.

4. Seed

Someone connected to the network that holds a complete copy of the file being shared.

5. Peer

Someone connected to the network that does not hold a complete copy of the file being shared.

The downloading process

Downloading a file using BitTorrent occurs in two stages. For the first stage, you will need to get the small .torrent file that points to the download you want, and there are plenty of sites where these can be found online. By opening the .torrent file (either from the site itself or from your computer after downloading it), you will activate your client and begin the second stage (the main download process).

The second stage starts with the client connecting to the tracker and acquiring the list of users currently sharing the file you want to download. These users consist of seeds and peers, and together are referred to as a swarm. A BitTorrent download cannot begin without a tracker, but once connected, the client would be able to complete the download should the tracker become inactive.

All downloads are transferred piece by piece in a random order, and once all the pieces have been collected, the client can then reconstruct the file.

This is a slight disadvantage over traditional linear P2P systems, as you will have to download the entire file before you can use it, but it tends to download faster so it balances out.

When you enter the swarm for the first time (as a peer), your client will start to download a few pieces of the file at a relatively slow speed. After a short while the you will begin to share those pieces with other peers, allowing the transfer rate to increase.

When you've completed your download, you automatically become a seed by remaining online and continuing to share the file. The client will continue to 'seed' until it is closed or disconnected from the Internet.