Every few years, someone predicts the imminent collapse of the Internet. Bob Metcalfe–Ethernet inventor, 3Com founder, and Technology Review patron–famously said at a 1995 Web conference that he would eat his words from a pessimistic Infoworld column if the Web didn’t disappear within a year under the strain of traffic overloads and other problems. In April 1997, Metcalfe contritely drank a milkshake containing the torn-up bits of his column. In 2004, Helsinki University of Technology professor Hannu Kari said spam and viruses would kill off the Internet by 2006. The fact that you’re looking at this website now means Kari was wrong.

The point is that even the sharpest innovators and entrepreneurs have often had a hard time seeing how the next hurdle in the Internet’s growth would be overcome–only to be surprised by some new resource, technology, or business idea that emerges in the nick of time. (In both 1996-97 and 2005-2006, Internet service providers responded in part by adding more bandwidth to their networks.)

But will that pattern hold out forever? The dam-breaking success of YouTube and Apple’s iTunes Video Store–neither of which existed prior to 2005–has unleashed a huge new flow of digital video on the Internet, and many consumers now spend hours a day streaming or downloading everything from home movies to live sports and prime-time TV series. Because video files are so large compared with the Web pages and e-mail messages that used to dominate Internet traffic, backbone lines are under strain, and backbone operators such as AT&T and Verizon and Internet service providers such as Comcast are facing new costs they can’t easily recoup, given the flat-rate pricing of most consumer broadband Internet access plans.

Hui Zhang, a computer scientist at Carnegie Mellon University who studies broadband networks, says that “2006 will be remembered as the year of Internet video. Consumers have shown that they basically want unlimited access to the content owners’ video. But what if the entire Internet gets swamped in video traffic?”

This time around, the Internet may be saved by the unlikeliest of rescuers: the builders of peer-to-peer file-sharing networks. In the minds of many consumers–and many studio executives–P2P networks are still synonymous with digital piracy. After all, Napster, Kazaa, and other early peer-to-peer networks were playgrounds for copyright violators, who downloaded millions of music files they hadn’t paid for. But today a number of researchers and entrepreneurs are arguing that peer-to-peer technology–which allows network members to retrieve content by tapping into the hard drives of other members who have already downloaded that content–is also great for distributing legitimately purchased, copyright-protected music and video. It might even lessen the burden on service providers and content distributors.

By putting new twists on the P2P concept, businesses hope to bring digital distribution of movies, TV shows, and other premium content to a market beyond iPod owners and YouTube addicts. In August 2005, for example, Internet company Wurld Media, of Saratoga Springs, NY, rolled out a P2P platform called Peer Impact. Once users have downloaded the free Peer Impact media player, they can buy and download movies and TV shows (as well as games, radio shows, and music) to a shared folder on their PC. The twist: Peer Impact pays users every time someone else on the network draws that content from the shared folder. Members can use their “Peer Cash” to buy more content. (I tried Peer Impact’s service recently and found downloads to be fast and smooth. I was able to start watching a one-hour episode of the fanfic production “Star Trek: New Voyages” about two minutes after the download process began; the entire download took about twelve minutes.)

Even BitTorrent, an advanced P2P network long seen by movie and record executives as an irksome successor to Napster and Kazaa, is going mainstream. The company announced early this month that it had raised $20 million in a second round of venture-capital funding and acquired competitor μTorrent (pronounced “microtorrent”), maker of a compact version of the BitTorrent software meant to be suitable for set-top boxes and other non-PC devices. BitTorrent–which speeds up downloads by grabbing and reassembling file fragments from the most accessible peers on the network, rather than by transferring whole files from one peer to another–is still one of the best tools for locating and procuring Internet video. That’s in part because it’s free and in part because so many people use it and have built a worldwide archive of digital files.

Many other companies are jumping into the P2P video mix, including Peercast, Octoshape, Allcast, and Itiva. Dijjer, a project of Revver (which was founded by peer-to-peer pioneer Ian Clarke), reduces the load on individual users’ computers by grabbing much of the content of a requested file from other users’ computers. A team of researchers from three universities in the Netherlands has introduced Tribler, a BitTorrent-like program that adds useful features such as Amazon-like recommendations and real-time maps showing who is downloading the same content. And in the United Kingdom, the BBC is developing a peer-to-peer media player called iPlayer. In a trial conducted with 5,000 users from November 2005 to February 2006, users were able to download first-run BBC shows for seven days after their official TV broadcast; many participants used the service simply to catch up on their favorite programs, but interestingly, the network also found that study participants were disproportionately drawn to several new shows and “niche” shows that hadn’t fared as well among regular broadcast viewers.

The Internet is already teeming with peer-to-peer traffic. In fact, P2P downloads may account for as much as 60 percent of network traffic–and as much as 60 percent of that traffic is video, according to CacheLogic, which has developed a proprietary system for accelerating P2P downloads. (The system enhances P2P distribution by giving peer-to-peer networks access to dedicated, high-capacity “edge servers” scattered around the Internet, in much the same way as the traditional content-distribution networks pioneered by companies like Akamai.)

So how could additional P2P traffic actually be a good thing for the Internet? Carnegie Mellon’s Zhang points out that because peer-to-peer networks exploit both the downlink and uplink capacities of users’ Internet connections, they distribute content more efficiently than centralized “unicast” technologies. Zhang also says it should be possible to label P2P traffic so that service providers can track it and decide how much of it to allow through their networks. He and colleagues from the University of California at Berkeley have founded a startup, Rinera, to develop software that will give service providers such control.

“The network itself needs to be informed about the types of traffic it’s handling, and service providers need to participate by setting policies,” says Zhang. “Otherwise, as applications like video downloading really take off, we will see a congested network, which will in turn impede the development of video-sharing technology.”