(p.485) APPENDIX B How a Telephone Works and the Basics of Telecommunications and Networking
(p.485) APPENDIX B How a Telephone Works and the Basics of Telecommunications and Networking
There are two fundamental reasons for having a basic understanding of the technology underpinning telephony. First, the infrastructure comprising a network is complex and massive and thus is a major driver, along with the regulatory agencies, in the construct of most portions of the telecommunications world and its practices. Second, there has been a fundamental shift in the base technologies of telephony from analog to digital over the last two decades of the twentieth century. These have important implications for the its future. The changes are as profound as, say, moving from making ships out of wood to metal. Of course people cannot often actually see the difference in telephonic technology, but they know the difference in the clarity of calls and the functions they can perform with the digital versus analog. Most people think of the telephone as the device on their desk or mounted on a kitchen wall. In the Telecommunications Industry that device is called a handset or terminal. It is merely the tip of the iceberg, tip of the pencil, the starting or ending point of a long chain.
Think of a telephone as part of a system, a network of many technologies. This network includes telephones, transmission media, call-and message-switching equipment, and signaling technologies that control the traffic of voice and data from one point to another. All telephone communication systems have essentially four basic elements: the telephone itself; the transmission equipment that carries a signal (voice or data) from the handset into the network; switching hardware and software; and human operators, who used to receive calls and switch them to lines that direct the message to their intended destinations but now are more often troubleshooters. These all involve various techniques for connecting one telephone (p.486) phone circuit to another. Coordinating all these is signaling, which is all about how telephone networks are controlled and instructed to connect specific calls.
The basic concept of the telephone is that a handset converts the sound waves produced by someone's voice into analogous electrical waves, which are then transmitted over wire to the intended recipient, via the four parts of the network listed above, and then converted back into sound that can be heard by the listener. Traditional telephones have carbon granules through which an electrical current flowed, with sound creating pressures on the granules using a thin diaphragm, thereby affecting the intensity of the electrical current flowing through the handset out over the telephone line; newer phones used digital means to do the work of carbon granules. At the opposite end the reverse occurs. Switching involves connecting the signal transmitted by a speaker into the line toward the intended recipient. Originally, a telephone operator did this by physically plugging one line into a switchboard to connect to that person. That process was continuously automated all through the twentieth century; first, by using precomputer technologies, later computers, and by the late 1980s fully digital technologies. AT&T devoted decades and hundreds of millions of dollars to improving switches. In the process, it produced a vast body of telephonic technology.1 Local control offices were located all over the nation, connecting local calls to transmission lines that allowed conversations to go long distance to other local control offices, which then passed calls to intended recipients. Over the decades the capacity of the nation's telephone network to handle more calls rose nearly exponentially.
In the “old” world of telecommunications (pre-1984), the switch was the key value component of the network and was the property of telecommunications companies; businesses and homes had telephone lines, but none had a switch. The switch has a long history dating back to the mid-nineteenth century and was the innovative technology that allowed telecommunications as a business to evolve from telegraphy to telecommunications. A century later, when switching became digital, switching could be handled by various proprietary systems, and the key value component of the telecom business began evaporating. Today, physical transport and switching are rapidly being commoditized, raising serious questions about how telecommunications companies can add value to the network now that their more than century-long key technology—the switch—is no longer under their sole control.
So far, I have described what is normally called analog signals. In other words, the voice is converted to electrical impulses—known as the voice signal—and has the same oscilloscope readings as the original signal (the voice). By the late twentieth century, analog also came to mean transmissions that were digital. When implementation of digital telephony became possible, the voice went into the phone as before—they still are analog signals—but then was translated into digital impulses (the binary code of digital signals) wherein the voice is sampled many times a second and each sample is assigned a binary series of numbers, through a process called modulation. The digital signals now look like standard computer-based digital data, and are carried through the line, switches, and so forth, to the intended recipient.
(p.487) Digital switching (also called packet switching) was first tried in 1962, and by 1969 over a half-million voice channels were in use across the nation. Digital telephony enjoys two fundamental advantages over prior approaches. First, signals can be reproduced exactly as originated, as opposed to analog transmissions, which lose signal strength and pick up distortions over long distances which is why long-distance telephone calls were often hard to hear and scratchy until late in the twentieth century. Second, digital circuitry is much less expensive to install and maintain than the pre-digital circuitry, which consisted of a great mass of copper wires. In the digital era, optical cables can carry many times more calls in a fraction of the number of lines. Furthermore, digital circuits can be fully automated and controlled by computers. The one great negative in using digital is the cost of converting from analog systems, an expense borne by American telephone companies.
There are other considerations with the digital to keep in mind. Once a telephone company converted sound to digital signals, then it could also transmit data (such as pictures, text) over telephone lines, because the technology required to convert text to digital forms (the on/off electrical impulses in a computer) had been used for decades as the central technology of computing. Since data that could be read by a computer were electrical impulses and electricity flows through a telephone line, one could now send all types of information over a telephone line. So data, voice, pictures, and so forth, were converted to digital signals, transmitted to the recipient, and then converted back into images or sound so that the person could understand the messages.2 Once that was possible, then the name of the game was to speed up the transmissions and increase the capacity of the telephone systems to handle ever larger amounts of data over a wire. Improving modulation techniques was important and more feasible in a digital world. Prior to the arrival of digital telephony, existing technologies were more or less adequate to handle the volume of telephone calls, but with crude data transmission starting in the 1960s, the pressure to handle greater volumes more quickly drove technologists toward the digital.
We should recognize that there is a vast difference between optimizing a network for voice calls versus data calls. In a voice network—such as what existed for a century—the intelligence to manage the flow of conversations lay in the network. These are called “smart networks” because switching technologies control the flow of events. In a data network, intelligence for managing the flow of information resides in the data, which contains information, and in the devices and applications that package the data and tells it where to go. A data network does not need a traffic policeman or stoplight as might have existed in a traditional analog switch. It also does not matter for data networks whether the data travels via main roads, highways, or back roads as long as the data reaches its destination. In voice networks, however, all traffic is carried on main roads and highways because these kinds of networks segment their bandwidth into dedicated lines or time slots, which is akin to having a limit on the number of lanes and cars per lane. In the digital scenario, data traffic can always get to where it is going on time.
(p.488) We should understand the effects of the Internet on telephony. In the early days, one connected a PC to the Internet by dialing a telephone number using an existing telephone network, going through an existing network of switches. Data transmissions tend to carry more electrical signals on a line (data never pauses the way we do in conversation; it only slows down if the line is crowded with other data), resulting in enormous traffic jams on the network. Eventually, the length of time users were connected live on the network exceeded the amount of time of a typical conversation. As time went by, users simply acquired a second telephone line dedicated to Internet access, which is why by the mid-1990s we saw a dramatic increase in the volume of new telephone numbers and area codes. So the hunt for more capacity and speed was on. Part of the capacity could be handled by the use of optical fiber, of which massive amounts were installed across the nation, but also in speeding up the transmissions themselves, hence the 1990s phrase, “broadband.” Specific speeds of transmission determined if a line was a narrowband (able to handle one voice channel, as in old POTS), wideband, which was faster and used to move data in the 1980s and 1990s, or broadband, which handled even more traffic (to be technical, capable of handling more than T-1 rates, or roughly 24 channels at wideband speeds).
It is important to understand the role of increasing demand for telephone numbers and area codes. This happened first because of growing demand (described in Chapter 7), but also because government regulations established zones to make sure the country would not run out of telephone numbers. The government also sells blocks of telephone numbers to carriers that they can issue within their zones. These two regulatory practices have a tremendous impact on local and regional switching because each switch needs to be programmed to recognize the zone from which the call originates and the zone to which it is being directed. Therefore, the system for handing off calls from zones to regions to national networks has become dramatically more difficult to do over time. All switches had to be reprogrammed, particularly in the 1990s.
One way to speed up transmission speed was by taking signals (a conversation) and separating them into bunches (packets), then shipping these down a line with other packets of partial conversations—much like highway traffic, which is all intent on moving things and people in the same direction. The ability to ship pieces of transmissions, called packet switching, grew out of digital technologies. In effect, packet switching made it possible to keep lines fully loaded with transmissions but organized so that all the pieces of, for example, your conversation, could be put back together in a timely and correct manner at the other end of the line. That meant one could simultaneously use the same line at home both to take a telephone call and to log onto the Internet. This ability was the true genius of the innovations that Vinton Cerf and Paul Baran had developed in the early 1960s. Add in the hunt for faster transmission speeds and one begins to see the continued complexity of the telephone network in the new century.
Finally, let's discuss the lines themselves. Twisted pair, which is the copper telephone wires that have been used for a century and continue to be deployed in every home and office, can handle about 7.7 million bits of information per second, (p.489) or roughly 120 telephone conversations. Fiber optics can handle 45 million bits of information per second, or just over 700 telephone conversations. By using packet-switching approaches to transmit waves of light through such networks (called Wavelength Division Multiplexing, or simply WDM), one could, as of 1996, push through a line some 40 billion bits of information per second, which translated into 625,000 telephone conversations. By 2000, that combination of WDM and fiber optics could handle 1.6 trillion bits of data per second, or 28 million conversations. Now one can understand why the telephone companies in the United States made vast investments in optical networks. Between 1996 and 2000, installation of fiber optic capacity grew at a 158 percent compound rate.
In Chapter 6, I noted that as the century progressed, data transmissions became an increasingly important part of the traffic over a telephone line. Then came widespread availability of the Internet, which is based on digital technology. So we need to understand further the differences between the pre- and post-Internet telephonic technologies. Prior to the arrival of the Internet, switching consisted of a body of technologies called circuit switching, which dated back to the first decade of the twentieth century (albeit with improvements over time) and was the basis of POTS. Two devices connected through a dedicated line on the network (called the circuit), and voice or data moved back and forth in the same order as originally transmitted. Consumption of bandwidth was fixed, in other words, the two parties owned the entire capacity of the line between them regardless of whether they used all its capacity or not; nobody else could use the line. The origins of this approach date back to the days of telegraphy. When telephony first came into use, dedicated wires were strung from town to town, and each line was a party line, so one had to schedule specific times to talk. Since the original concept of voice switching was based on this concept, it survived through the years even though technologies now afford us more effective ways to create a channel of communications between two distant parties. Digital approaches represented a radical departure from the old model.
In the world of packet switching, digital is now the basis of the Internet and most data communications services. Software manages the networks and determines routes. One's use of a telephone line varies, since there can be other packets of data also using the same line, so dedicated access is never guaranteed. Of course, if a network were down, packets can bypass a bad line and arrive at their destinations via alternative paths.3
In the 1990s, a technical debate, tantamount to a holy war, was in full swing in the Telecommunications Industry. To put it bluntly, on one side were the traditional telephone industry representatives, pejoratively called “Bell Heads,” and on the other side the “Net Heads.” The Bell Heads argued that switching networks needed to be highly intelligent, that is, smart enough to know that one wanted a telephone line and who to get the call to. They argued that the telephone itself should be “dumb” with no processing capability, leaving computing intelligence in the network itself. Thus, the network is the service of value to the user. The Net Heads argued the opposite, specifically that networks should be dumb and that the telephone (or terminal) should be smart and have computing capability. (p.490) In their model, networks are simply utilities that, like the street in front of one's home, merely provide facilities for “smart” drivers to direct their vehicles wherever they want to go. As the Internet became a preferred channel for handling the movement of data, activities, voice, and sound in the 1990s, one could quickly realize how important such debates had become.
Telephone and other telecommunication provider companies are built around one view or the other. Many of the tens of thousands of Internet Service Providers (ISPs), so prevalent in the 1990s, sided with the Net Heads, while the Bells were trapped into maintaining preexisting, older technologies and infrastructures. In the early 2000s, as more traditional telephone and cable companies began rapidly to displace ISPs in the fast-growing market for high-speed Internet connections and telephone services, one could again quickly see the debate taking on an urgency. The business dilemma all this presents is that no service provider has been able to figure out a viable business model to transition from one operating model to another. Nor has any determined when it makes most sense operationally to transition—that is, when one can make the most money or create the most value for end users. Meanwhile, traditional phone systems continue their historic transformation to fully digital, high-speed, packet-switching approaches, leaving behind long-standing technologies of the old Bell System. As with so many other technologies, therefore, we see that how a telephone works is as much a discussion about business issues as it is about technology, electronics, and physics.4 Telephony and its industry were originally about physics, then electronics and technology, but now are all about fully developed businesses, and that governs the other three issues.
(1.) A. Michael Noll, Introduction to Telephones and Telephone Systems, 3d ed. (Boston, Mass.: Artech House, 1998): 1–125; Stephen M. Walters, The New Telephony (Upper Saddle River, N.J.: Prentice-Hall PTR, 2002): 15–107; for pure analog systems, see David Talley, Basic Telephone Switching Systems (New York: Hayden, 1969). The body of knowledge about digital telephony was large even by the time of AT&T's breakup and had been codified and widely available in the industry; see, B. E. Keiser and F. Strange, Digital Telephony and Network Integration (New York: Van Nostrand Reinhold Company, 1985).
(2.) Because of the flexibility of digital technology, voice and data appeared the same on a line and thus even the FCC did not track the differences between one kind of traffic or another, per conversation between Jim Lande of the Industry Analysis Division of the FCC and the author, October 21, 2003. For the kind of tracking the FCC does, see Linda Blake and Jim Lande, Trends in the International Telecommunications Industry (Washington, D.C.: Federal Communications Commission, April 21, 2001). The only tracking was done mainly by various government agencies and ISPs interested in the amount of traffic going through the Internet.
(3.) Walters, The New Telephony, 43, 48, 50; Jane Laino, The Telecom Handbook: Understanding Business Telecommunications Systems and Services, 4th ed. (New York: CMP Books, 2002): 19–20.
(4.) Susan E. McMaster, The Telecommunications Industry (Westport, Conn.: Greenwood Press, 2002): 153–175.