Just as his company has begun to put large businesses onto its well-publicised new data network for all kinds of communications -- from transferring files to making voice calls -- Esrey this morning told a gathering of industry leaders at The Wall Street Journal Technology Summit that it's time to prepare for life after the Internet.
"The Internet is going to be replaced by other nets. The Internet is going away, as we know it," he said. "There will be a rationalisation of economics." At issue: The charges that long-distance phone companies pay for access to local phone networks when they want to complete phone calls for customers.
Currently, such access charges amount to about 45 percent of the cost of long-distance calls, Esrey said. Meanwhile, with government policy bent toward fostering the development of the Internet "as we know it," data calls face no such charges, even if the packets actually are carrying voices. "It makes no sense," Esrey said, because both voice and data calls will use the same amount of resources, as communications companies move toward some kind of packet-based network as Sprint did with its new Integrated On-Demand Network. In his mind, the "day of reckoning is right around the corner".
Robert Metcalfe, inventor of Ethernet packet networking technology and vice president of technology at International Data Group, disputed Esrey. "That idea is really not on [the mark]," he said after Esrey's talk. Instead, he envisions "five to 10 next-generation networks," based on different forms of exchanging packets at high-speed. Some of those networks are expected to be based wholly on Internet protocols. However, communications experts such as Bell Communications Research Inc. Chairman Emeritus George Heilmeier believe that, in the future, they also will handle such packet technologies as frame relay and Asynchronous Transfer Mode. "This tends to confuse telephone people," Metcalfe said, "so the answer [to them] is, the Internet is dead."
Data customers naturally will resist the addition of access charges for Internet services, said John M. McQuillan, president of McQuillan Consulting. That will put phone companies such as Sprint in a bind. Sprint, Esrey noted, still derives half of its profit from local phone services. If the Net starts handling significant amounts of traffic that otherwise would be conducted as local phone calls, access charges likely will have to be reduced, McQuillan said. This will squeeze the bottom lines of conventional phone companies. "That's not the Internet's problem," he said. "That's the local telephone company's problem."
Another key aspect of the Internet "as we know it" also will change over the next few years, Esrey said. Concerns about the capacity of the Internet and other data networks to handle rapidly escalating demand will diminish. Instead, the new bottleneck, he contended, will be software applications and application servers. Companies will not be able to effectively handle the downpour of packets that will result from the emergence of big communication pipes in data networks.
In 1989, 24,000 simultaneous phone calls could be conducted on a single pair of optical fibers. In 1998, that number is escalating to 2 million simultaneous calls. Next year, it will reach 33 million, Esrey said.
Esrey is missing the point by "defining the Net in terms of the number of phone calls" it can execute at one time, said Douglas E. Van Houweling, president and chief executive officer of the University Corporation for Advanced Internet Development. The organisation is spearheading the development of academia's next-generation Internet 2 technologies. Instead, Van Houweling said, he should be focused on videoconferencing, messaging and other business applications that will dominate the new internetworking of communications and computing. "That's the net of the future," he said. "That's the Net where we're headed.