Mbones and Giganets

FOR SOMETHING THAT HAS SPREAD WITH ALL the forethought of kudzu, the Internet isn't half bad. Unless you care that only about 10 percent of Americans can access it, that the equipment to tap into it can cost upwards of $1,000, that it's as easy to navigate as the Northwest Passage, that voice and video come through at glacial speed or not at all ... the inadequacies go on. How much better a network could you create if you actually planned it? That vision inspires dreamers and schemers who are planning the Nextnet. Tapping the new networks, they say, might require little more than a phone and TV; sound, video and data would zip through something as prosaic as power lines. "We have in the Internet the seeds of this [next network]," says Vint Cerf, a founder of what became the Internet and now a senior vice president at MCI. This time, "what people want, and what people do," will shape what comes next.

What do people want? Services like voice and video--the Internet delivers them only slightly faster than snailmail. To provide this, the Nextnet--actually, it's unnamed so far, but Giganet and MBone and National Information Infrastructure (NII) are in the running, and you can guess which snappy label is the government's--would have info pipelines whose bandwidth swamps today's standard phone lines. One contender, the Gigabit Testbed Initiative, envisions billion-bit-per-second wires. Ways to connect and manage such capacity are now being tested at five sites across the country. Or the signals could be carried by phone lines or TV cables. But there's a dark horse. Utilities are thinking of installing "information pipes" into homes to carry electronic signals that would control peak power demand (users wouldn't notice the slight brownouts). "They might just as well put in high-capacity telecommunications fiber, because they could charge for services on it," says Cerf

High voice-data-video capacity needs smart technology to route it. The Multicast Backbone, or MBone, is software that now routes videoconferences over the Internet (though you need powerful workstations to make it work). Upgraded, it could carry video signals to simple PCs, explains co-inventor Steve Deering of Xerox Palo Alto Research Center. A system of electronic switches called Asynchronous Transfer Mode (ATM) already runs a few local networks, and routes data faster and with less loss of integrity than on most of the Internet. "The big question is whether ATM will replace the Internet," says Richard Binder, who is working on the gigabit network project.

With such agile software, the Nextnet could offer better hypertext links to tame the chaos of the World Wide Web, and download video and software long before flying toasters arrive.

The Clinton administration, which wants to make the Infobahn accessible to everyone, calls its vision of the wired future the National Information Infrastructure. Spearheaded by the Commerce Department, this "network of networks" is already forming. Some of the 92 projects that received $67 million in NII grants last October establish the kind of local networks that, linked nationally, begin to look like a Nextnet. An Alaska network will connect 81 percent of the population; an Iowa project already links government agencies, schools and libraries with optical fiber (NEWSWEEK, Dec. 19,1994); 30 "freenets" will provide free Internet access but could also be the backbone of its successor. Expect wireless hookups, says Cerf, with signals bouncing from transmitters like cell-phone relays. Or, tomorrow's "information appliance" could be a TV, computer and phone all in one. Even cheaper, a device similar to an adaptor could feed Nextnet data into a regular TV No telling what you can net with forethought.