First generation P2P file sharing

These networks, the most famous being Napster, were built around a 'centralized server'. Users would store the files they wanted to share on their own hard disks and run the Napster software, or 'client'. This would connect their machine to Napster's central servers and upload a list of the files they were sharing, turning each user's machine into a mini file-server, with the Napster server being the hub.

So Napster now had a complete master list of every file available to share on every hard disk connected to the central server. When a user ran a search query on the title of a song or whatever they were looking for, Napster would look for that title in the central database sending back a list of all the positive matches. The user would then select which one they wanted to download, and the client would connect the user's computer to the machine (peer) that had the file.

The problem with this approach was the centralised server that held the list of all the file titles. Eventually the courts decided this was in breach of copyright and forced the site to shut down, disabling the network.

Second generation P2P file sharing

A new system called Gnutella was developed to solve the problem of the centralised server. It works in the same way as Napster in that the users store the files they want to share on their own computer but it uses a decentralised structure and so has no central server to shut down. The network is therefore made up purely of users' machines that are running a Gnutella client and these machines are commonly referred to as nodes.

So when you begin, the Gnutella client connects your machine to just two or three nodes. These nodes will in turn be connected to two or three others and so on until you are connected, directly or indirectly, to many thousands, several layers deep.

When you run a file search, the request is passed along through the network and any positive matches reported back to you along with the host's IP address. When you select a file to download, your client automatically connects you to that host for direct downloading of the file.

Over the years, Gnutella has been developed and improved with the implementation of various features to increase download success rate and speed. One of the most significant is the creation of the Supernode.

Also known as an Ultrapeer, these are machines on the network with more bandwidth and processing power than the average user. Each time a new client connects to the network, it sends a list of all the files that user is sharing to the Supernodes, which have all the search queries routed through them to greatly shorten the time it takes to return positive matches.

Third generation P2P file sharing

This type of network functions in the same way as the second generation P2P file sharing networks do in that they are decentralised, but they also have anonymity features incorporated into them. They are essentially private networks that protect your privacy while you are connected, masking your IP address, encrypting what you send and receive, and rendering you untraceable.

As yet though, these networks have not achieved mass usage where file sharing is concerned due to the overheads required for the anonymity features that make them slow and hard to use.

Third generation networks:

Freenet

Entropy

GNUnet

I2P

A further evolution of the anonymous network is the Friend-to-Friend network, where only known users or 'friends' can connect to your computer. The search requests and file downloading is all passed along through the chain of nodes, with no direct connection between you and the file's host. This ensures that your IP address is only shared with people you trust, ensuring that it remains secure.

Friend-to-Friend networks:

Ants P2P

MUTE

WASTE