Let’s start with an overview of the key concepts of BitTorrent architecture to gain a high-level understanding of what we’re dealing with.

General concepts

BitTorrent is a protocol for peer-to-peer file sharing. The idea behind peer-to-peer file sharing is that there’s no single central location where the file is stored. Instead, multiple copies of the same file are located at many nodes (peers) in the network, and each peer can serve the contents of that file to others. Of course, it all starts with just a single node having a single copy of the file: the seed. But as more peers download and host that file, they can also start functioning as seeds themselves.

In addition, the file being distributed is divided into segments called pieces. Peers exchange and store pieces between each other, so even if a peer doesn’t have the complete file yet, it can act as a source for the pieces it has already downloaded, thus actively participating in the data exchange.

This scheme offers several advantages:

  • Increased download speed: peers usually download pieces of files from multiple locations in parallel.
  • Increased availability: as the file spreads across the network, more and more peers can seed that file.
  • More balanced bandwidth usage: since file pieces are downloaded from multiple locations, this relieves the original seed from having to serve them to everyone single-handedly.

Torrent file

Information about distributed files is published in the form of torrent files. A torrent file contains important metadata about the published file, such as:

  • The URL of a torrent tracker (announce URL) that clients initially use to discover peers.
  • Information about pieces: the size of a piece in bytes, and a cryptographic hash of each individual piece (SHA-1 or SHA-256).

Torrent trackers

There is one centralized node in the whole architecture, called a torrent tracker. Its main purpose is to assist in peer discovery. When a client initiates a file download, it connects to the tracker using the announce URL from the torrent file and obtains a list of peers that can serve the requested file. After that, peer-to-peer communication can continue without a connection to the tracker. Most clients, however, still communicate with the tracker periodically to provide network performance and download statistics.

Since the necessity for a central torrent tracker contradicts the decentralized nature of peer-to-peer communication, a few other methods have emerged to avoid this bottleneck: distributed hash tables (DHT) and the peer exchange protocol allow for peer discovery without dedicated torrent trackers.