Peer-to-peer in practice

A new book on peer-to-peer technologies that spends just five pages on copyright issues seems a bit counterintuitive. After all, copyright battles have propelled peer-to-peer implementations such as Napster and Gnutella to the front pages countless times. But editor Andy Oram's achievement in the new book, "Peer-to-Peer: Harnessing the Power of Disruptive Technologies," is a collection of sociological and technical wri-ting that addresses the potential of peer-to-peer without becoming trapped in intellectual property law.

The book is important not because it is well-written, or because it is completely up-to-date (impossible for any book about the Internet that also involves ongoing policy shifts), but because it provides technologists with an intelligent discussion about an emerging social phenomenon. On the other hand, its audience extends beyond designers. In fact, technically literate policy-makers might learn more from the details than those already steeped in system design jargon.

Typical definitions of peer-to-peer technologies rely solely on their decentralized nature, but several authors debunk this definition. For example, the quintessential new peer-to-peer technology, Napster, connects a decentralized group of users through a centralized file locator. In his chapter in "Peer-to-Peer," Tim O'Reilly points out that the difference between decentralized and centralized systems are not black and white and that different people may have slightly differing viewpoints.

One interesting definition suggests that peer-to-peer technologies are those that route around the traditional Domain Name System to help individuals interact directly with others. This definition of peer-to-peer includes instant messaging and expands the mainstream view.

Dan Bricklin's chapter on the "Cornucopia of the Commons" is a succinct and straight-forward description of what has made peer-to-peer technologies so successful. Bricklin argues that the genius of Napster is not in its file structure, which is awkward at best, but in its ability to force users to share their own existing files to gain new ones from other users. By stopping free rides on its system, Napster has become a true community of individuals.

Meanwhile, others say peer-to-peer technologies are not immune to the tragedy of the commons that strikes so many decentralized systems. After all, many government agencies, universities, businesses and Internet service providers have banned Napster simply because of the bandwidth concerns posed by users interacting directly.

The following chapters detail the lessons learned from some of the more important peer-to-peer technologies, such as SETI@Home, Gnutella, Freenet and Publius. The book concludes with recommendations for addressing policy issues that arise from peer-to-peer technologies. Privacy, trust, accessibility,, security and accountability are just a few concerns accounted for — all of great interest to those trying to develop new systems or looking to develop policies for using existing technologies.

Schwartz is a policy analyst at the Center for Democracy and Technology in Washington, D.C.