Public Switched Telephone Network
The Public Switched Telephone Network (PSTN) is, as its most essential function, the service that can establish connections among the world's telephones that permit generally open access to one another. As a practical matter, it also encompasses the technical standards and operating procedures that both the public telephone network, and specialized nonpublic (e.g., military) telephone networks operate.
When telephones first were invented, they were interconnected by individual proprietary networks. A well-connected office, at the turn of the twentieth century, might have several telephones, each operated by a different company. As an early U.S. example, two of the larger telephone companies were Bell and Home; a Bell subscriber could not call a Home-only subscriber, and vice versa.
As telephone use became widespread in large companies, and then international, it became obvious that no single organization would control all the telephones. It was a practical necessity that telephone organizations, many of which were initially government-operated or regulated monopolies, would have to agree on a common set of technical standards, so different operating agencies could interconnect with one another. In the United States, the 1913 Kingsbury Compromise granted a long-distance technical monopoly in the interest of what Theodore Vail, chief executive of the American Telephone and Telegraph (AT&T) corporation called "universal service". Vail's concept was not one of universality to individual users, but that there could be universal connectivity among separately managed telephone systems. The Compromise required AT&T to give up the right to offer telegraph service, that monopoly primarily going to Western Union. Perhaps defying logic, AT&T did not rename itself American Telephone.
In much of the world, however, telephone service remained a government monopoly, often under a derivative of the postal service, usually called a "PTT" for the French abbreviation for "Ministry of Posts and Telegraphs". Commercial networks such as AT&T, which was a descendant of the Bell System, were relatively rare until the late twentieth century. The PTT-dominated international standards bodies, which carried out the necessary function of defining technical specifications among different telephone systems, referred to the commercial systems as registered private operating agencies (RPOA).
While the first telephone operating networks used human operators to make connections even within the same company, it was apparent this could not increase indefinitely, as there would not be enough people to manage the connections. The telephone industry became the first to use a significant degree of automation.
The PSTN needs to maintain compatibility, even if it requires converters at the entry points to the network, with low-technology Plain Old Telephone Service.
While modern systems tend to refer to 4 kHz analog channels, most of the important energy to understand speech is distributed between 300 and 2800 Hz.
Analog vs. digital end instruments
Until the middle or later part of the twentieth century, telephones generated analog signals. From roughly the 1960s onwards, the signal might be converted, either at a private branch exchange at a multi-tenant building, or at the first telephone end office, to digital signals used in the transmission system.
Conversion to digital signals, still in a circuit-switched transmission system, became more common with the advent of the Integrated Systems Digital Network (ISDN) and proprietary equivalents in the 1970s. Digitizing, at the telephone, into a packetized format came into practice with the advent of voice over Internet Protocol (VoIP) in the 1990s. The nominal voice channel of ISDN used 64 kilobits of bandwidth, with relatively simple conversions between analog and digital signals. Certain military voice systems, limited by the bandwidth of tactical radio, gave up speech quality to gain more channel capacity.
Eventually, however, the mass production of speech processing microcircuitry allowed more efficient coding and decoding (i.e., using codecs), with pure speech channels, of acceptable quality, needing as little as 2.4 Kbps or less. Such codecs exploited the statistical properties of human speech, so they might not, for example, pass facsimile or data modem signals.
ISDN and other digital techniques did offer advantages for call control, especially supplemental services such as teleconferencing or caller identification. They also lent themselves to integrated voice with computer data applications.
Analog vs. digital transmission
Until electronics were produced in mass quantities, only analog transmission was possible. Early electronics, however, still sent and multiplexed analog signals, until the early 1960s. For a number of years after the first commercial digital transmission systems, it was still common to see purely analog signals between the telephone and either the PBX or end office.
Even after the advent of digital transmission, long-range transmission tended to use analog multiplexing over coaxial cable or microwave radio. This trend began to change with the availability of digital optical transmission systems in the 1970s.
The beginnings of automatic switching
Separating control and voice paths
As with the first transmission facilities, the first telephone transmission systems were optimized for analog signals, generally sent over copper wire. The most basic distance limitation on such transmission is simple electrical resistance, with a non-electronic telephone being able to drive no more than thousands of feet/low kilometers.
The next limitation was that certain analog frequencies would transmit more efficiently than one another, causing an impairment variously called attenuation distortion or frequency distortion. Among the first workarounds to this problem was counterintuitive to the problem of resistance-based limits. Inductors, called loading coils, were placed along the physical line, most commonly every 6000 wire feet in North American systems. Loading coils attentuate high-frequency signals, but have the effect of keeping the overall frequency response faithful to the frequency vs. energy distribution in speech.
Scaling the transmission facilities
Putting each call onto a separate copper pair became mechanically impractical. Since the frequency response of even basic copper wire is greater than the range needed for an individual speech path, it became common practice to modulate calls onto radio frequencies sent down the wire, providing "pair gain" or analog multiplexing. For relatively short distances, analog multiplexing technology tended to focus on groups of 12 voice paths; longer-range systems, where the cost of transmission media, was of even greater concern, used multiples of groups. One common long-haul technique built mastergroups of 600 paths (i.e., 50 groups), and some systems carried multiple mastergroups.
In circuit-switched digital telephony, the technologies available tended to multiplex, at the first level, into digroups of 24 (North America) or 32 (Europe) calls. Since the early digital technology had limited range, the conversion from local digital to long-haul analog tended to take place at a granularity of 600 calls.