Peer-to-Peer

Peer-to-peer (P2P) interaction refers to the ability of any two information appliances, such as computers, to connect directly with one another without going through any central intermediary.

The notion of P2P interaction is not unique to the Internet. Consider the following scenario. Someone who is interested in obtaining a recipe for strawberry rhubarb pie can pursue a couple of strategies: go to a library (or log onto the Web) and search a central repository of documents for a recipe, or ask one or more friends if they have a recipe. A person who followed the first strategy would be a “client” obtaining information from a “server” that stored all of the information at one location—the library or Web site. A person who followed the second strategy would be engaging in P2P interaction, bypassing any intermediary central collection of recipes. For any given purpose, it may be far more effective to pursue one type of strategy or the other.

The same two strategies are in effect on the Internet. The advent of the Web made it possible for some large Web sites to “serve” information to millions of “clients” who used Web browsers. Users can find recipes, buy books, and read the news on many well-known Web sites, such as http://Amazon.com, Yahoo! and AOL. More recently, there have been many applications on the Internet that also make it possible for users to pursue P2P strategies. These applications fall into three broad categories: file sharing, computer-resource sharing, and communication and collaboration.

First, there are applications that allow peers to share information (audio, text, video, and graphics files) directly with one another without going to a Web site. Perhaps the most popular such application was Napster, which allowed individuals to share music files on their computers with millions of their peers anywhere in the world. Some have argued that Napster is not a “true” P2P application, because it had an intermediary “server” that provided a directory of which peers had what songs on their computers. There are other applications, such as Gnutella and Freenet, that dispense with the [Page 368]intermediary directory server and are therefore indisputably P2P. With these applications, a user would send his request for information (a song or any other digital document) to a small subset of peers who were connected to the application; they in turn relayed the request to a further subset of their peers. Soon, the user would have a list of peers from around the world that had the information that she requested.

Second, there are applications aimed at sharing computing resources among peers' computers, rather than sharing files. Clay Shirky, an expert on Internet applications, observes that “At a conservative estimate … the world's Net-connected PCs presently host an aggregate of 10 billion megahertz of processing power and 10 thousand terabytes of storage.” In many cases, these Internet-connected PCs have spare storage space on their hard drives. In addition, these PCs are not being used much of the time. In 1995, scientists at University of California–Berkeley proposed using the computing power of these PCs to help in the search for extraterrestrial civilizations.

...