In the ever-evolving world of the internet and the economic imperatives that drive it, the essential truths still apply: The more things change, the more the past becomes the future.
In the mind of visionary Vinton Cerf, the future of Internet access is less likely to be ruled by the panoply of gadgets that require points and clicks, and more likely to involve the oldest form of intelligent communication: the human voice.
Cerf — known worldwide as the legitimate father of the Internet for designing its basic architecture nearly 30 years ago — says that speech, in concert with devices that can convert voice commands to data and distribute it in other forms, may well replace the clackety-clack of the keyboard.
And that could mean that the most important device for accessing the Net might well be the one globally deployed gadget: the telephone.
“I dont know what the economic impact of this might be,” Cerf says, “but I am beginning to wonder about the interesting developments in speech recognition, and how it opens up another way of interacting with the Net: through the telephone.
“We are just starting to explore speech-enabled Web sites. Im not sure if that will create a huge demand or not. But the billion or so telephone devices out there today, already in place, become ways to interact with the Internet. I think that will have a salient effect” on the improvement of voice recognition.
“One thing starting to show up on my horizon is interaction with Internet appliances, making it possible to dictate what will come out as text. What one might have happen is you pick up a phone, an ordinary phone, and call a speech-understanding machine, and the result might be conveyed to you in other ways. It might go to a display in a car, or to initiate a transaction in some other medium,” Cerf says.
Convergence, or cross-communication between devices, is an important part of what Cerf sees as the future of Internet use, especially that which turns speech into a tool to drive Internet-connected appliances, like the refrigerator, the garage door or monitoring systems in the home and workplace, a la George Jetson.
But with these opportunities come significant and possibly difficult challenges that must be addressed over the next five years to make clear what Cerf describes as a “business-sensible” approach to these technological advances.
In an Internet economy littered with failed business models — what Cerf calls the fallout “from the dot-bombs” — there are serious challenges that go far beyond the problems of last-mile interconnectivity and whether Web advertising can keep the Internet within economic reach of the average consumer.
For one thing, as the number of Internet-enabled devices has boomed, the Web has fast begun to run out of Internet Protocol (IP) addresses. Over the next three years, Cerf says, the explosion in mobile devices will make that an increasingly critical problem.
That issue has already led to the development of IP version 6, which Cerf helped shape. IPv6 solves the problem of dwindling numbers of addresses and improves routing and network autoconfiguration. But it brings with it the problem of integrating the protocol in use today, IPv4, with the new standard.
“There are a lot of details yet to work out,” Cerf says. “A lot of administrative systems will have to be upgraded . . . and all standard applications will have to be modified to use the [newer] protocol, to make the new systems and old systems work with each other. And we need to get busy on that now.”
Though build outs of optical fiber networks continue, Cerf is concerned that bandwidth demand will exceed availability, stalling the use of fat-pipe applications like downloading video.
“There are a lot of applications that just wont work without enough bandwidth,” he says. “If we can get a Gigabit Ethernet up and running at the end of the Net . . . it would allow us to deliver an enormous amount of information in a short amount of time. But the only way to do that is to get fiber closer and closer to the destination.”
And Cerf worries that DSL service, which is being pushed as a solution to high-speed access needs, will eventually disappoint both consumers and providers.
“DSL may eventually turn out to be more expensive to maintain as a physical plant than optical fiber,” he says. “So economics may drive us in the direction of fiber to the home.”
But that continues to raise right-of-way issues, which Cerf says causes him to look toward electric utilities for a solution.
“Very few people have been building fiber to residences, but when you look at the power companies, especially in California, you have to wonder if they need another business line. In addition to supplying electricity, they might permit the use or right of way to carry photons as well as electrons.
“Optical fiber is relatively immune to interference, as is typical power cable, because they operate in the same regime. It might be that power companies could become the dark horses of the fiber business,” he says.
Cerf also likes what he sees in the commingling of Internet access and global positioning systems that allow computers to identify a users location.
“If a cell phone knows where it is, and can send that information on the Net, you could speak to a computer that understands voice, and it could deliver the data you need, based on your geographical location — like the best restaurants [nearby].”
Beyond that, Cerf, who is senior vice president for Internet architecture and technology at WorldCom; chairman of the Internet Corporation for Assigned Names and Numbers and a member of international forums, says unresolved language issues will continue to be a barrier to the best development of a global Internet.
“There has been strong lobbying by Asian speakers, but also by people who use something other than the Latin alphabet who would like to enter the equivalent of, say, IBM.com. But to do that, for instance, in Russian, there are unbelievably complicated problems with what seems like such a simple thing.
“The system now doesnt have a good way of saying, This came in Russian or German or Japanese. Farsi and Arabic character sets look the same. Character sets cross languages. It is very complex,” he says.
But Cerf seems to thrive on complex problems. In 1972, he and Robert Kahn, a colleague from the University of California at Los Angeles, developed the Transport Control Protocol/Internet Protocol technology that allows computers to talk to each other.
And today hes working with with the National Aeronautics and Space Administration in the development of an interplanetary IP that would facilitate communications between Earth, spacecraft and, someday perhaps, colonies on other planets.