When you think of peer-to-peer (P2P) computing, what immediately comes to mind? Do you think of two computers communicating with each other? Do you think of the online music-sharing powerhouse Napster.com, the site that got into legal trouble by providing one of the first music-sharing services on the Internet? With the pervasive deployment of computers, P2P is not only receiving increasing attention in research, product development, and investment circles (not to mention the hundreds of thousands of P2P participants already using this type of distributed computing technology), but is also the basis of the Second International Workshop on Databases, Information Systems, and Peer-to-Peer Computing, held in Toronto, Canada in August 2004.

The authors present a number of papers on P2P-related topics. So what is P2P? P2P refers to a class of systems and applications that employ distributed resources to perform a function in a decentralized manner. P2P computing isn’t really new; the fundamental technology behind P2P has been around for a while in the form of completely decentralized networks of peers.

The authors suggest that a P2P computer network topology consists of a network that relies primarily on the computing power and bandwidth of the participants in the network. P2P networks are typically used for connecting nodes largely via ad hoc connections. Such networks are useful for many purposes. In the purest sense, a P2P network does not have any notion of clients or servers, but uses the notion of peer-node equality, where each simultaneously functions as both a client and a server to other nodes on the network.

The authors state that the P2P approach has a number of benefits, such as improving scalability by avoiding dependency on centralized points; eliminating the need for costly infrastructure by enabling direct communication among clients; and enabling resource aggregation. In fact, P2P computing promises to be the prime paradigm shift necessary to push a number of interesting distributed computing technologies out of the shadows and into the light.

Interestingly, the problems that still need to be solved in P2P computing loosely map to problems encountered in distributed computing, such as coordinating and monitoring the activities of independent nodes and ensuring robust, reliable communication between nodes. However, not all distributed computing is P2P computing.

The authors do not propose a single P2P design solution in this volume, but the P2P practitioner may want to use the authors’ research findings as a basis for consideration when investigating P2P further.