From LQWiki

Peer to Peer networking (often abbreviated as "P2P") is, as the name may suggest, about communicating amongst equals. More rigid computer networks are typically client/server based, relying on a few heavy duty systems to control and distribute various forms of information (from IP addresses to documents) to each individual machine. The most widely known of these are probably web browsers and web servers This form of network works well in an established and maintainable environment but does not suit all the needs of network users.

Without these central systems, each system stands as itself talking to the other networked systems freely. The easily recognizable form of a peer to peer networks are likely to be found in file sharing networks, such as Napster and Gnutella; however such a network can also be found in a huge number of small businesses around the world using adhoc IT solutions. For example, take a small architects' site, with a handful of Macs and people who don't wear shoes. Without a central location to store all their drawings, plans, and other documents, they keep them on their own machines, and share certain folders on their system so that other people can access the files. This, too, is a peer to peer network, as no one system (or group of systems) is in charge.

There are also varying extents to which a given network can be seen to operate in a P2P fashion, with many such examples falling between two stools, as it were. A good example here, as mentioned above, can be found in the ill-fated Napster. The main cause of its downfall is the fact that the users and shared data were held on a set of central servers dotted around the Internet. Searches and queries were done against the logs of available files on those servers. This part of the operation is clearly client/server based, however once these results were obtained and a user wished to download a file, it was transferred directly between the two clients, hence making it actually a P2P network. If Napster did not have these central servers, it would quite probably be bigger and badder now than it was at its peak.

Peer to peer is often referred to as a disruptive technology. In a client/server environment, the client will generally always know where the server is. With nothing resembling a server in existence within the network, other peers are discovered from broadcast requests and other such techniques which, if not continually optimized and improved, can lead to a much heavier network load, and a loss of service.

For an entertaining an in-depth read about what peer to peer networking is, the O'Reilly book Peer-to-Peer: Harnessing the Power of Disruptive Technologies[1] is a great read.

P2P Applications

For actual P2P applications, please see File Sharing