OpenCola isn't a soft drink company. It offers a free recipe for brewing cola in a bathtub, but that's just a marketing gimmick. The company's primary business is selling so-called open-source software for Web sites, which allows customers with similar interests to link up and search each other's computers for information on whatever excites them. OpenCola believes auction and content sites will be so impressed by these peer-to-peer (P2P) connections they'll buy $100 million worth of its software by 2004.
Only problem is, while OpenCola's software is widely acclaimed, its sales projection looks fizzy. The company may be a microcosm of the P2P software sector: It is still struggling to bag its first customer.
A year ago, with Napster still in full swing, companies large and small touted P2P as the next big thing -- not only for consumers but for businesses. Its promise attracted dozens of startups aiming to sell P2P systems for everything from searching corporate networks to helping workers in separate locations collaborate on projects.
REMEMBER THE BOOM? Bear Stearns analyst Chris Kwak estimates venture capitalists invested about $500 million in hundreds of peer-to-peer companies during last year's P2P boom (remember it?). Then the tech sector imploded -- and with it went P2P's near-term prospects. In fact, with P2P software companies having so much trouble raising funds, experts expect serious consolidation soon.
That should work to the advantage of big fish such as Microsoft, Sun, and Intel, which also have plopped into the P2P pond. Yet in a time of slowing spending on information technology, even they will have trouble persuading companies to adopt systems that radically change their business practices. The technology "isn't really the barrier," says Rob Batchelder, an analyst with tech consultancy Gartner Group. "It's more the business model," he says, explaining that no one has figured out how to make money on many uses of P2P.
For big companies, the promise of P2P is threefold. By using idle PCs, P2P connections could tap into cheap, underutilized processing power and bandwidth using a technique called distributed processing. By linking users directly, P2P could create easier collaboration, allowing the rapid formation of work groups that sidestep traditional barriers such as firewalls and restricted intranets. Finally, P2P could make information more accessible throughout an enterprise by opening up the desktops of individual employees -- allowing staffers to search each other's virtual desk drawers more freely.
PIECEMEAL PROCESSING. For now, most of the companies pursuing P2P seem to be focusing on the less-nosy variety, distributed processing. In this application, a network of interconnected PCs would process large chunks of data and combine to perform tasks once reserved for supercomputers.
Each PC would get a small bit of data, process it, and send it back to where it originated. Then, the results from all the PCs would be reassembled to complete a job. Although it's slower than using big iron, as mainframes are known, distributed processing should be significantly cheaper because companies could use their existing fleets of PCs.
To date, the proof of this concept has come mainly from the nonprofit world. Since 1998, University of California at Berkeley's Professor David Anderson has recruited 3.2 million volunteers to each download a small program that parses radio telescope data as part of a massive search for extraterrestrial life. The program runs while the volunteers' PCs are idle and sends the processed data back to Anderson every time they log on. This network has the processing power of 3.2 of IBM's $100 million ASCI White supercomputers -- muscle that Anderson got for free.
Anderson's experiment underlies new features being marketed by United Devices -- an Austin (Tex.) company that creates distributed processing networks for corporations -- and competitors such as DataSynapse and Entropia. Of all the P2P business possibilities, distributed processing "is the smartest [investment] bet near-term," says Will Porteous, an associate at RRE Ventures, a venture firm that has invested $6 million in San Diego-based Entropia.
GIVE UP CONTROL? Another possible application of P2P is so-called "enterprise collaboration" software. Such software lets numerous users work on the same project in real time and create working groups on the fly. Thus, it mirrors the way business increasingly functions today, with a mix of outside contractors and company staffers coming together to work on projects, says Ray Ozzie, creator of the original collaboration software, Lotus Notes.
To reach the market, Ozzie in 1997 launched Groove Networks. Using Groove's software, employees can share files, mark up virtual whiteboards, and surf the Net together in a shared virtual space. Groove's product even passes through corporate firewalls and directly connects groups of PCs. That way, users can set up groups without hassling network administrators for special permissions. As for security, Ozzie built the system to fully encrypt activity, cloaking information from hackers.
Sounds good, except IT managers dislike relinquishing control over information flow within their organizations, and fear security risks from P2P connections. Thus, software aimed at changing business processes to the extent P2P would has become an increasingly tough sell. Says IDC senior research analyst Robert Mahowled: "I don't see corporations in large numbers giving up the server soon, especially in these cautious times."
That said, a handful of big companies are testing P2P collaboration systems, including networking giant Cisco Systems, defense contractor Raytheon, Ford Motor Co., and the U.S. Defense Dept. One customer of collaboration software maker NextPage reported that productivity among employees who used its software rose more than 25% in less than a year, says Darren Lee, NextPage vice-president of strategy and marketing. "We already know that people are willing to spend money on this," he adds.
UNEARTHING TREASURES. OpenCola falls under the third P2P prospect in the business realm, the building of distributed search and sharing capabilities. The company's flagship product, called Folders, would let visitors to auction sites watch what other shoppers are doing and buying. OpenCola Chief Evangelist (yes, that's his official title) Cory Doctorow -- an avid collector of Disney memorabilia -- uses the software to see what Disney items other collectors are bidding for, and at what price.
Also known as content sharing, such systems are designed to function on an enterprise-wide scale. An estimated 66% of data never leaves individual PCs to make its way to corporate servers and storage, according to Bear Stearns. So P2P content distribution could unearth treasures and unlock intellectual firepower by allowing each network user to see every file on the network, or at least a larger portion of those files.
Of course, such systems might not prove as useful as P2P proponents claim. Many of the files stored on individual desktops are personal or just plan useless to anyone except the PC's owner. Merely finding something helpful on another user's computer might not be enough. "I guarantee they don't want to see notes, they want to talk to the person," says Andrew Mahon, director of strategic marketing at Groove.
THE BIG BOYS P2P. That's not keeping the industry giants out of the fray. Microsoft, server-maker Sun, and chip-maker Intel all have P2P projects in the works. In partnership with Groove, Microsoft recently introduced its premier P2P program, HailStorm, which combines Microsoft Passport technology (which holds your name, credit card numbers, and the like) with collaboration features. For instance, if a HailStorm user books a flight through an online travel agency, the program would pay for it with that person's credit card.
Sun in April launched its Project JXTA, offering an open-code P2P platform able to interact with other peer devices. So far, developers have downloaded 100,000 copies. Sun is already working with 40 companies to explore various uses for the product, says Ingrid Van Den Hoogen, director of marketing and business development at JXTA. "We are in the very early stages of people trying it out," she adds. "The killer apps are going to still emerge."
If so, part of the reason could be P2P security protection unveiled by Intel in February. The Peer-to-Peer Trusted Library helps with peer identification and encryption. And last August, Intel organized the Peer-to-Peer Working Group, which is striving to come up with industry standards for P2P software. Member companies include OpenCola, NextPage, and Groove.
FACT OR FICTION? Killer app or not, P2P software isn't expected to gather much money until 2003, says Gartner's Batchelder. When it comes to distributed computing, companies remain somewhat leery of turning big projects over to PCs. With no single platform or protocol, meanwhile, companies risk stumbling into a P2P Tower of Babel when it comes to collaboration. Indeed, users would likely adopt collaboration software only if they can talk to those both inside and outside their organization, according to Simon Yates, an analyst with technology consultancy Forrester Research.
Still, the P2P crew keeps chugging along. OpenCola's Doctorow, an award-winning science fiction writer, hopes an unpublished novel he recently penned presages a P2P future. His Down and Out in The Magic Kingdom portrays a future in which OpenCola's own Folders product lets people use information about others to help manage their lives -- with all the advantages and disadvantages that implies. The question now is whether P2P can make the leap from the world of fiction to real life. By Olga Kharif in New York