|
| P2P. The term conjures up images of subversive teenagers on Napster downloading free music behind closed bedroom doors. Never mind that many of Napster's 60 million users are downloading classical music, not Eminem, and have teenagers of their own. The fact is, P2P--shorthand for peer-to-peer computing--is much more than music-swapping.
It's using the processing power of millions of individual PCs to search for extra-terrestrial intelligence with the SETI@home project or to explore AIDS treatments with the FightAIDS@Home initiative. Peer-to-peer is also about companies tapping their employees' PCs to crunch numbers or share knowledge. And it's dozens of other innovative applications that have yet to hit the market. Already, some $300 billion in venture capital has been invested in P2P startups, and the number of outfits calling themselves P2P companies has swelled to nearly 150. ``What venture capitalists look for most is a disruptive technology that changes the competitive landscape,'' says Steve Brotman, a partner at Silicon Alley Venture Partners. P2P, he adds, ``is more disruptive than Netscape'' and its introduction of the Web browser.
So what exactly is it? Part architecture, part philosophy, P2P is a style of distributed computing that exploits the resources of far-flung computer nodes on a network. Instead of being mostly passive slaves, or ``clients,'' to all-powerful computer servers, each connected device becomes a fully participating peer in the network. These peers bypass the central control points of traditional computing systems, shifting power to the individual nodes. If computer architecture were politics--and some maintain it is--then P2P systems would be libertarian, in contrast to the authoritarian systems of client-server, or browser-Web server.
P2P is not a new idea by any means. Indeed, the original design of the Internet was to let any computer node talk directly to another without an intermediary. What has changed is the power of those computer nodes. Today's PC is more powerful than a circa-1980 Cray supercomputer. P2P expert Clay Shirky at the Accelerator Group, a New York incubator, figures that the world's Net-connected PCs collectively pack some 10 billion Mhz of processing power and 10,000 terabytes of storage.
CHILD'S PLAY. Yet these resources are vastly underutilized. Think about your home PC: You might send some e-mail, surf the Net, compose a documentall child's play for your 300 Mhz Pentium 3. Most of the time, that computer is idle.
P2P unlocks the underused resources of these computers on the fringes of the Internet and allows them to be shared by a larger community. Napster creator Shawn Fanning's innovation was to distribute digital music files across the hard drives of his members' PCs rather than store them on a central server. In the SETI@Home project, 2.7 million computer users in 226 countries have donated the processing power of their PCs, creating a vast, distributed supercomputer to analyze radio signals from space for signs of alien life. And a company called Groove Networks Inc., founded by Lotus Notes creator Ray Ozzie and backed by $60 million from Accel Partners and others, is looking to help corporations unleash the knowledge locked away on workers' PCs.
One P2P project even aims to share phone lines. Called the Free World Dialup project, it takes Net phone calling a giant step further by allowing a person to ``borrow'' the phone line of someone else on the network and turn a long-distance call into a local one. That may sound like standard Internet telephony--but these are calls between telephones, not PCs. If you're in San Francisco and want to place a call to Helsinki, you just dial the number and the system routes it through the Net to another PC that is physically located in Helsinki. The Helsinki PC then places the local call.
SIGHT SHIFT. Applications like that give P2P its reputation as a subversive force. And they unnerve gatekeepers around the world--whether it's telephone companies who manage and route calls or record companies who control the pressing and distribution of music CDs.
But it's not all anarchy. The peer model can have real benefits for business. Just ask J.P. Morgan Chase & Co., which is using P2P to distribute big processing tasks, such as risk management calculations, across employees' computers at night. Or law firm Baker & McKenzie, which is implementing P2P software from startup NextPage Inc. in Salt Lake City to capture and share the knowledge of its 3,000 attorneys in 60 offices.
With any new architecture that changes the rules, there are opportunities for innovation. New search methods are needed to get at the previously invisible data and resources that P2P makes accessible. OpenCola and Infrasearch Inc. are two companies working on that problem. Infrasearch, a Burlingame (Calif.) startup doing real-time distributed search and backed by Netscape co-founder Marc Andreessen among others, was acquired in early March by Sun Microsystems Inc. XDegrees, based in Mountain View, Calif., offers a central directory for peer-based resources that companies can use to set up P2P systems. Aimster, a startup in Troy, N.Y., is creating applications that can piggyback onto the instant-messaging infrastructure. And Centrata Inc., a venture funded by Kleiner Perkins Caufield & Byers in Silicon Valley, is still in stealth mode, but is building an enterprise computing platform based on P2P.
Proponents of this technology argue that the organic nature of P2P systems more closely reflect how businesses actually work. ``Organizations are evolving like complex adaptive systems,'' says Bob Andersen, a Groove Networks P2P evangelist. P2P facilitates spontaneous behavior, in contrast to software that supports structured, repeatable business processes (think payroll or purchasing systems). ``The P2P architecture will lead to entirely new organizational structures,'' says Paul Saffo, director of the Institute for the Future.
For these reasons, the knee-jerk reaction that P2P provokes in many established institutions, such as the recording industry, may prove short-sighted. Any new technology or idea worth its salt initially appears to challenge some segment. The VCR is a classic instance. Once seen as a threat to the film industry, the video business now accounts for roughly 70% of film revenues. Similarly, P2P networks like Napster may actually be a boon to the music industry--once artists and record companies figure out new business models. Indeed, surveys show that Napster use may boost CD sales.
The P2P architecture will not displace the centrally managed systems that handle much of today's business and commerce. But it's already clear that P2P is here to stay. No matter what the fate of Napster is, other file-sharing systems such as Gnutella and Freenet are taking hold. There's even a project under way to add many common Internet protocols--for things like e-mail, news and chat--to Freenet, which could create a parallel Internet that is completely anonymous and impossible to shut down. Like the Internet itself, P2P is an agent of change--and probably for the best. |
Copyright 2001, Bloomberg L.P.
Printer-Friendly Version