1 2 3 4 5 6

roughly 3 percent. The $350 million infrastructure upgrade AOL announced in January 1997, for completion in July 1997, would take the service only up to the lower end of that industry standard. Love's argument was that even the high end of ISPs' network provision, 10 percent, was lower than the one in seven Bell Atlantic was claiming.

Settlement is not the only pricing option that has been proposed. Hal Varian, an economist at the University of California at Berkeley, has been arguing for some years that the Net needs priority-based pricing, which would give higher-quality service to those willing to pay for it. In a sense, this type of service is arriving already with the private, value-added networks being built by companies like IBM and GEIS, which sell data network services at a higher price, promising better security and faster transmission in return. These services route data over their own networks as far as possible, handing over to the Internet, if necessary, at the last possible moment. But Varian points out that this structure of alternative networks doesn't suit the home-based consumer, who ideally should have a choice of quality of service over a single line.[13]

Bill Washburn, the founder of the Commercial Internet Exchange, an industry body usually known as CIX set up in 1991 to promote the commercialization of the Internet, believes a lot of the talk of doom is misplaced. "I think this is just a recent variation on the Net Is Dead comment--or the Net Is Dying or the Net Is About to Die--that they've been saying for years. It's a tradition--this will kill the Net."[14]He believes that the current pricing structure has been a vital one: "Flat- rate pricing made new kinds of activities and services possible and freed the imagination."

Nonetheless, it may be too easy for Netheads to trust that whatever application they come up with will eventually attract the bandwidth it needs by some kind of divine mandate, and to underestimate the risks of placing all their faith in what is arguably a Heath Robinson contraption.

As Stephenson traced the his way through the Middle East and the wilds of Cornwall, he studied the map of international connections and concluded: "Netheads have heard so much puffery about the robust nature of the Internet and its amazing ability to route around obstacles that they frequently have a grossly inflated conception of how many routes packets can take between continents and how much bandwidth those routes can carry." In mid-1996, the entire state of Minnesota was cut off from the rest of the Net, he notes, when its single primary connection went down. "If Minnesota, of all places, is so vulnerable, one can imagine how tenuous many international links must be."[15]