In Peer-to-Peer: Harnessing the Power of Disruptive Technologies, proponents of peer-to-peer systems offer insights into the history, makeup, and technology behind these networks. The book was published in early 2001, at a time when the term “peer-to-peer” was creeping into the headlines of mainstream news. The book also came out just months before the landmark court ruling against Napster in July of that year. As many of the examples used in Peer-to-Peer center around Napster—the most prominent peer-to-peer networking system at the time—and as many of the key players in the peer-to-peer world have changed in the past six years, the book can seem dated. However, the popularity of file-sharing and the legal issues surrounding peer-to-peer did not die with the Napster court case. So while Peer-to-Peer may seem dated, it is still relevant, as it provides readers with a valuable understanding of how peer-to-peer works offering a backdrop to the current issues facing these networks.
Peer-To-Peer is separated into three sections: Part I: Context and Overview; Part II: Projects; and Part III: Technical Topics. The first section helps define the makeup peer-to-peer technology, as well as its history. One important theme rises out of these first four chapters: building peer-to-peer systems, and the motivation behind them, are not new trends.
For example, in “A Network of Peers: Peer-to-Peer Models Through the History of the Internet,” authors Nelson Minar and Marc Hedlund point out that the Internet, as it was originally devised, was a peer-to-peer system, and that this history offers examples to draw from for designing future peer-to-peer systems. It wasn’t until the rise of the World Wide Web and the Web browser in the mid 1990s that the client/server model rose to dominance. In chapter three, Tim O’Reilly adds that even the Web started out as a participatory groupware system, and that “Only later was it recast into a publishing medium, in which sites seek to produce content that attracts millions of passive customers” (pg. 53). As millions of users joined the Internet at this time, a slew of new issues popped up, such as spam and security tools like firewalls, which eroded the “open network” required for peer-to-peer systems to thrive.
Part II of Peer-to-Peer, “Projects,” takes a closer look at a sample of peer-to-peer systems and applications. These examples include Jabber, Gnutella, Mixmaster Remailers, Freenet, and SETI@home. Some of these examples are described by their creators, and in turn the description of benefits to these tools sometimes turn into hyperbole. In the chapter describing Gnutella, developer Gene Kan calls the system “a landscape-altering technology in and of itself. Gnutella turned every academically correct notion of computer science on its head …” (pg. 94-95). Regardless of whether this statement is true or not, one begins to wonder how much of what they’re reading is marketing material. Still, the chapters in this section help show how peer-to-peer networks can come in different forms, as well as ways in which they disrupt the client/server model. In the same chapter on Gnutella, Gene Kan playfully compares Gnutella’s decentralized system to a cocktail party where you ask your friends sitting next to you to pass the sushi tray (i.e., specific files), and then uses the same analogy with Napster, in which the partygoer must locate the party’s host to help find the sushi tray, or the specific files he or she wants. Through such examples we begin to see the differences in how peer-to-peer networks can be organized.
Through the example networks in the second section of Peer-to-Peer, we also see how these networks offer advantages over the traditional client/server Internet model. Freenet addresses issues of privacy; Publius combats censorship; SETI@home taps into the collective network to allow supercomputing for scientific research. Some of the examples, such as Publius and Free Haven, have very similar goals, but overall this section reveals the reach and variety of peer-to-peer networks. Those who still see peer-to-peer as simply a tool for sharing mp3s and video will be surprised to learn how many peer-to-peer systems are being used.
The final section of the book—“Technical Topics,”—tackles the architecture and design of peer-to-peer technology. “Metadata” discusses the challenges and possible solutions to searching for information on peer-to-peer networks when different data sources aren’t connected, unlike the ease of using a browser on the Web. “Performance” asks how decentralized peer-to-peer systems can create more reliable networks, even as they rely on users’ computers to remain available and accessible. Also, peer-to-peer faces performance issues under increased traffic.
“Trust,” “Accountability,” and “Reputation” discuss the risks that come with sharing information over peer-to-peer networks. In “Trust,” Marc Waldman and Avi Rubin show how successful systems will be designed to require as little trust as possible, so “there is no uncertainty about how the system will behave. In this context, the ideal trusted system is one that everyone has confidence in because they do not have to trust it” (pg. 245). “Accountability” and “Reputation” cover ways to restrict access to unfavorable users and attackers, as well as building ties with favorable users through reputation servers with scoring algorithms – tools that can rate the quality of a peer by harnessing feedback from other users.
It is in these last chapters and in some of the examples from Part II that the book may show its age. It is not that the material is necessarily obsolete, but more recent technological developments in peer-to-peer design, as well as the inclusion of new popular peer-to-peer networks would make the material much more relevant.
This past week, a Comcast client filed a lawsuit against the company for blocking access to peer-to-peer sites like BitTorrent and Gnutella, claiming that the company is using software to hinder and even cancel file-sharing. In the same week, Warner Music CEO Edgar Bronfman gave a speech admitting that the companies’ aggressive actions against music file-sharing was a mistake, as they refused to believe that they needed to change their business model and, in turn, were too slow in adapting to consumer tastes. Undoubtedly, the rise of peer-to-peer systems is far from reaching a plateau, and years from now we may look back to many of the authors in Peer-to-Peer as visionaries who changed the landscape of the Internet. Gene Kan may have been right after all.
