P2P may rise in Napster's wake

The days of free music may soon be over, but peer-to-peer file swapping could be the next big thing.

COMMENTARY--Peer-to-peer (P2P) content distribution may be the biggest fad to hit computing since the explosion of the Web. Popularized by the infamous Napster, P2P file sharing lets users' PCs act as Internet file servers.

In Napster's case, the music-sharing service provides only a directory. Someone looking for a particular file would search the service's directory, then Napster would simply provide a pointer that lets the seeker load the file directly from the source PC at no charge. Because most of these files were copy

righted songs, the record companies were successful in challenging the site's legality. Even though Napster's servers never touched the copyrighted files, the directory gave Napster enough information to be pretty sure the files listed were copyrighted. Users would have been better off if the files were stored on Napster's servers rather than on users' PCs. That, however, would have been too blatant a copyright violation.

Legal issues aside, Napster's technical advantage was that its servers required relatively little storage and far less bandwidth than if it hosted the files locally. This is the enduring advantage of P2P file sharing. P2P lets all the computers on a network act as one massive file server, potentially delivering the cumulative storage and bandwidth of all the PCs on the network. This can be especially useful for distributing huge multimedia files to large numbers of users; each PC that receives the file can act as a server for additional PCs.

P2P possibilities
The problem with P2P file sharing, at least for consumers, is there just aren't that many files people want to share. Plus, each PC must remain on and connected to the Internet for its files to be available, and the transfer speed is dictated by the speed of the source PC. With the free exchange of copyrighted material no longer a totally legal option, it's unclear how important P2P file sharing will be for consumers. It could, however, give rise to some valuable business applications, from enabling business partners to streamline their joint business processes, to allowing companies to distribute video information to many employees quickly and without massive server bandwidth.

P2P technology could also apply to any interaction where one user gets information directly from another user's computer, without explicitly accessing the Web. A central server may act as an intermediary, in which case the interaction isn't technically P2P. The server can store information in transit, so the sending and receiving computers don't have to be on and connected at the same time. E-mail is today's primary example of this kind of interaction. In the future, we'll see many other uses of this approach, which is sometimes called asynchronous, or mediated, P2P.

Another major application of P2P technology is the sharing of computing resources rather than files, an approach more accurately called distributed computing. The best-known example of this is SETI@home, an initiative to analyze a massive body of data collected by the Arecibo radio telescope to search for signs of extraterrestrial life. The data set was sliced up into manageable chunks users could download along with the analysis code so their PCs could analyze it. Although the effort didn't find any signs of intelligent life (extraterrestrial, that is), it did manage to enlist more than 2 million PCs in the cause.

United Devices, a startup seeking to commercialize this technology, has contracted with Exodus Communications, a major Internet hosting company, to use the network of consumers' PCs to perform Web-site stress testing. More in the SETI@home vein, it also has a project under way to enlist millions of PCs in a computational chemistry effort to search for a cancer cure.

The SETI@home and cancer-cure efforts are essentially philanthropic causes that have a natural appeal. But will consumers be equally willing to offer their computers in the service of Exodus' for-profit enterprise? I doubt it. And if United Devices were to offer even a few dollars in compensation—hardly a major motivator—it would cost the company many millions of dollars, putting the potential profitability of the enterprise in question.

The Internet provides an infrastructure that enables direct communication among computers on a global level—a remarkable technical accomplishment that's at the heart of P2P. Ultimately, P2P is just another set of tools in the Internet toolbox. But the extreme degree of sharing it allows could raise social, ethical, and business issues beyond what we've already seen with the shutting down of Napster. It's these issues, rather than the technical ones, that will be instrumental in determining P2P's future.