Because the Internet is inexpensive to use, people assume it is government funded. That isn’t so. However, the Internet is an outgrowth of a 1960s government project: the ARPANET, as it was called, was initially used solely for computer-science and engineering projects. It became a vital communications link among far-flung project collaborators but was virtually unknown to outsiders.
In 1989, the U.S. government decided to stop funding ARPANET, and plans were laid for a commercial successor, to be called the “Internet.” The name was derived from that of the underlying communications protocol. Even when it became a commercial service, the Internet’s first customers were mostly scientists at universities and companies in the computer industry, who used it for exchanging e-mail.
The financial model that allows the Internet to be so suspiciously cheap is actually one of its most interesting aspects. If you use a telephone today, you expect to be charged for time and distance. Businesses that call one remote site a great deal avoid these charges by getting a leased line, a special-purpose telephone line dedicated to calls between the two sites. There are no traffic charges on a leased line—the same amount is charged for it each month no matter how much it is used.
The foundation of the Internet consists of a bunch of these leased lines connected by switching systems that route data. The long-distance Internet connections are provided in the United States by five companies, each of which leases lines from telecommunications carriers. Since the breakup of AT&T, the charges for leased lines have become very competitive. Because the volume of traffic on the Internet is so large, these five companies qualify for the lowest possible rates—which means they carry enormous bandwidth quite inexpensively.
The term “bandwidth” deserves further explanation. As I said, it refers to the speed at which a line can carry information to connected devices. The bandwidth depends, in part, on the technology used to transmit and receive the information. Telephone networks are designed for two-way private connections with low bandwidth. Telephones are analog devices that communicate with the telephone company’s equipment by means of fluctuating currents—analogs of the sounds of voices. When an analog signal is digitized by a long-distance telephone company, the resulting digital signal contains about 64,000 bits of information per second.
The coaxial cables used to carry cable television broadcasts have much higher bandwidth potential than standard telephone wires because they have to be able to carry higher-frequency video signals. Cable TV systems today, however, don’t transmit bits; they use analog technology to transmit thirty to seventy-five channels of video. Coaxial cable can easily carry hundreds of millions or even a billion bits per second, but new switches will have to be added to allow them to support digital-information transmission. A long-distance fiber-optic cable that carries 1.7 billion bits of information from one repeater station (something like an amplifier) to another has sufficient bandwidth for 25,000 simultaneous telephone conversations. The number of possible conversations rises significantly if the conversations are compressed by removing redundant information, such as the pauses between words and sentences, so that each conversation consumes fewer bits.
Most businesses use a special kind of telephone line to connect to the Internet. It is called a T-1 line and carries 1.5 million bits per second, which is relatively high bandwidth. Subscribers pay the local phone company a monthly charge for the T-1 line (which moves their data to the nearest Internet access point) and then pay a flat rate of about $20,000 a year to the company connecting them to the Internet. That yearly charge, based on the capacity of the connection, or “on ramp,” covers all of their Internet usage whether they use the Internet constantly or never use it at all, and whether their Internet traffic goes a few miles or across the globe. The sum of these payments funds the entire Internet network.
This works because the costs are based on paying for capacity, and the pricing has simply followed. It would require a lot of technology and effort for the carriers to keep track of time and distance. Why should they bother if they can make a profit without having to? This pricing structure means that once a customer has an Internet connection there is no extra cost for extensive use, which encourages usage. Most individuals can’t afford to lease a T-1 line. To connect to the Internet, they contact a local on-line service provider. This is a company that has paid the $20,000 per year to connect via T-1 or other high-speed means to the Internet. Individuals use their regular phone lines to call the local service provider and it connects them to the Internet. A typical monthly charge is $20, for which you get twenty hours of prime-time usage.
Providing access to the Internet will become even more competitive in the next few years. Large phone companies around the world will enter the business. Prices will come down significantly. The on-line service companies such as CompuServe and America Online will be including Internet access as part of their charges. Over the next few years the Internet will improve and provide easy access, wide availability, a consistent user interface, easy navigation, and integration with other commercial on-line services.
One technical challenge still facing the Internet is how to handle “real-time” content—specifically audio (including voice) and video. The underlying technology of the Internet doesn’t guarantee that data will move from one point to another at a constant rate. The congestion on the network determines how quickly packets are sent. Various clever approaches do allow high-quality two-way audio and video to be delivered, but full audio and video support will require significant changes in the network and probably won’t be available for several years.
When these changes do happen, they will set up the Internet in direct competition with the phone companies’ voice networks. Their different pricing approaches will make the competition interesting to watch.
As the Internet is changing the way we pay for communication, it may also change how we pay for information. There are those who think the Internet has shown that information will be free, or largely so. Although a great deal of information, from NASA photos to bulletin board entries donated by users, will continue to be free, I believe the most attractive information, whether Hollywood movies or encyclopedic databases, will continue to be produced with profit in mind.
Software programs are a particular kind of information. There is a lot of free software on the Internet today, some of it quite useful. Often this is software written as a graduate-student project or at a government-funded lab. However, I think that the desire for quality, support, and comprehensiveness for a tool as important as software means that demand for commercial software will continue to grow. Already, many students and faculty members who wrote free software at universities are busy writing business plans for start-up companies to provide commercial versions of their software with more features. Software developers, both those who want to charge for their product and those who want to give it away, will have an easier time getting it distributed than they do now.
All of this bodes well for the future information highway. However, before it becomes a reality, a number of transitional technologies will be used to bring us new applications. While they will fall short of what will be possible once the full-bandwidth highway is available, they will be a step beyond what we can do now. These evolutionary advances are inexpensive enough to be cost-justified with applications that already work and have proven demand.