Click for next page ( 234


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 233
10 Combining Communications and Computing: Telematics Infrastructures DEAN GILLETTE Today's combinations of computers and telecommunication devices to form new infrastructures are as important to national economies in the twen- tieth century as the combination of steam engines and carts to form railroads in the nineteenth. The new infrastructures support the development of busi- nesses and industries of the "information society" businesses and industries that provide jobs for over half of the U.S. work forced Of all the names suggested for the new combination the most commonly used is telematics. It is the anglicized version of telematique, a word coined by Nora and Minc (1980) in their book L'Informatisation de la Societe.2 Telematics today is directly used in almost every sector of our socio- economic structure. For example, many people find automatic teller ma- chines convenient for personal financial transactions. A Wall Street Journal reporter types a story into a word processor in New York, and the paper is printed and published in Chicago, San Francisco, and Miami the same day as in New York. Automobile manufacturers tailor production sched- ules to fill dealers' orders for cars with specific arrangements of features. And the federal government relies on telematics for functions ranging from agency management to command and control of national defense systems, which themselves contain integrated communication and com- puter infrastructures. The technical and economic feasibility of such a range of telematics uses was achieved well over a decade ago. Newspapers and magazines, as well as learned journals, constantly describe new applications of te- lematics and growth in the variety and sophistication of old applications. 233

OCR for page 233
234 DEAN GILLETTE Much can be done with existing telematics components by imaginative applications and by building new infrastructures based on them. Even more will be possible as improvements are made in critical areas. One such area is the cost of high-speed data communication; another is the flexibility and cost of terminals. Fortunately, there is vigorous research and development in the underlying technologies, including solid-state elec- tronics, and advances in the field should lead to decreased costs and other improvements. A third area of concern is the lack of uniform standards for the interconnection of facilities and protocols for interaction between computers of different manufacture. Progress is being made in achieving these goals also, both domestically and internationally. The evolution and application of telematics, however, have also exacer- bated certain existing social problems and introduced new ones. Theft by telecommunication is now possible, and new threats to privacy have appeared. New dichotomies between rich and poor are growing domestically and in- temationally. Private managers and governmental bodies face new challenges in acquiring the benefits of telematics while mitigating its harms. This chapter reports on the nature and status of telematics infrastructures and on opportunities for improvements. It also touches briefly on prospects for future telematics systems and a few of the social issues such systems raise. COMPONENTS OF A TELEMATICS INFRASTRUCTURE Computer functions have been used in communications since about 1900, first with the introduction of the dial-controlled switch to make automatic telephone connections and, more obviously, since the 1950s when computers were attached to the telephone network to measure usage and computer charges. (The switching and billing computers, however, are internal to the network and are not accessible to the telephone user after a connection is made.) Communication lines carry information among processing, storage, control, input, and output functions in a computer. Indeed, transmission speed and data transmission capacity are critical, sometimes limiting, factors in designing advanced computer systems. The interaction of computing and communications in telematics differs from the use of computers in communications. In telematics, at least one terminal on a communications connection is a computer, and information is sent over the communications network in machine-sensible form, nor- mally in formats and at speeds completely incomprehensible to humans. Each telematics system has both physical and logical parts. The physical parts are the data transmission channels, the associated data switching systems, and the computers at terminals or nodes in the network. The

OCR for page 233
COMBINING COMMUNICATIONS AND COMPUTERS 235 logical parts are the standards and protocols for interconnecting and using the physical parts. In applications in which data processing is distributed, as with automatic teller machines in banking, the organization of this distribution is an aspect of the logical part. Except for such distributions, the computers that are part of a telematics infrastructure can operate equally well standing alone and not attached to data transmission facilities. To adapt the computers to telematics applications, two types of changes are needed: the computer must be physically adapted to treat the transmission facilities as devices for input and output, and the software must be prepared to permit interaction with the data network, terminals, and other computers on the network. Thus, from the vast infrastructure of computing, we may take as unique components of telematics the logic of standards and pro- tocols and the plan for geographical distribution of processing, both of which are to be represented in software. The applications programs of computing are not a part of the telematics infrastructure, but the infra- structure helps make their results widely available. Data Transmission Facilities The first electrical telecommunication system transmitted a digital sig- nal the Morse code of telegraphy. Today, although some communication facilities are limited to telegraph data transmission rates fire alarm and traffic light control networks, for example the vast bulk of communi- cation channels are designed to transmit voice signals. Because it is readily available and can be adapted to carry data signals, the technology for telephony is the foundation of the data transmission facilities in telematics. A component of a telematics infrastructure could be a voice-grade chan- nel in the telephone network. A pair of wires, a "loop," connecting a home or office to a local switching system network could be such a component. The channel could be a connection through the public switched telephone network or a private, dedicated voice channel in the network of facilities that make up the communications infrastructure. Almost any voice channel in the telephone network can transmit data at 1,200 bits per second (b/s), and 1,200-b/s modems cost less than $200. Although some pairs must be adapted, most wire pairs and all voice channels in carrier systems can carry data at a rate of 9.6 kilobits per second (kb/s). High-speed modems cost about $5,000. In addition to voice channels, the telecommunications infrastructure contains a hierarchy of facilities that transmit digital signals at various rates. At the bottom, or zero level, of the hierarchy, we find devices that encode a continuously varying voice signal into a 64-kb/s digital signal. At the first level of the hierarchy, 24 of these signals are "time division

OCR for page 233
236 DEAN GILLETTE TABLE 10-1 The U.S. Digital Transrrlission Hierarchy Level Data Rate Equivalent Telephone (Mb/s) Channels 0.064 o T1 T1A T2 T3 T4 1.5 3.1 6.3 44.7 274.2 24 48 96 672 4,032 multiplexed" into a single 1.5-megabit per second (Mb/s) stream.3 Forty- eight 64-kb/s signals are combined at level 1A. Table 10-1 shows the hierarchy of digital signal transmission in the United States. To begin to appreciate the potential of these existing facilities, note that the text of this chapter (about 5,000 words) could be transmitted in 0.1 s at the T1 level, and the text of all the chapters in this volume could be transmitted in 1 s at the T2 level. The signal transmitted over any transmission medium may be at any one of several of these levels. A pair of wires with regenerative repeaters spaced at 1-mile (ml) intervals can carry a T1, T1A, or T2 level signal. A coaxial cable or microwave radio relay system, either terrestrial or satellite borne, can carry a T4 level signal. An optical fiber system can carry at least two T4 level signals, and its capacity is increasing as research and development continue. Any digital channel in the telephone hierarchy can carry any data signal compatible with the maximum bit rate in that channel. Thus, instead of using a 64-kb/s data channel for a single voice signal, we might use it for six 9.6-kb/s data channels multiplexed in a digital logic of choice. A level 3 channel could either carry 672 voice channels or a properly encoded entertainment-grade television signal. Because the assignment of the meaning of the bits in any channel is arbitrary, these transparent channels of the communications infrastructure can be the components of a telematics infra- structure provided a protocol to define the meaning of a stream of bits has been established. Data Switching Facilities The function of switching is to choose the path a message will take while traversing a communications network. The message coming into a switching system must in some way contain, or be accompanied by, an

OCR for page 233
COMBINING COMAlUI!lICATIONS AND COMPUTERS ~23 7 instruction to the switching system regarding the outgoing channel a 10- digit telephone number, for example. Each switching system contains an internal connection network that establishes a path between an incoming and an outgoing transmission channel, and it contains a control function that responds to instructions to establish that path and order the connection network to do so. Over the years, three technologies have been developed and used in switching: electromechanical, electronic, and digital systems.4 Because almost all new transmission systems are digital, the choice of digital switches is economical in telephone systems and is natural in telematics. Telephone network switching systems connect one voice channel to another to establish a unique circuit between the calling and the called parties. At present, 5 s or more are needed to set up a cross-country connection through local and toll switches by circuit switching. But in 5 s, 50 pages of single-spaced text could be sent on a 1.5-Mb/s data link; thus, for some applications, the time involved in circuit switching is just too long. An alternative is to structure the network to operate like the postal service that is, to send the address of the receiver together with the message and have them forwarded together from switching center to switching center. In such a store-and-forward system, the entire message is passed from one node and stored at the next before transmission to the subsequent node. Alternatively, a long message may be divided into "packets" of specified length-say, 1,000 bytes each numbered and transmitted with the address of the recipient. In one protocol the packets are reassembled into the original message on reception. In another the first packet establishes a "virtual" circuit through the network for suc- cessive packets to follow. In addition to its speed advantage, packet switching is far superior to circuit switching for transmitting messages to multiple receivers; with a change of address at the head end of the packet, messages can be sent simultaneously to many locations. Terminals Terminals on a telematics network are referred to as "smart" or "dumb," depending on whether they have internal data processing capabilities; a microcomputer thus is a smart terminal, and a teletypewriter is dumb. Smart or dumb, the terminals must either generate a data signal in a format acceptable for transmission or connect to an interface unit that converts formats. Beyond the physical interconnection level, however, are the logical

OCR for page 233
238 DEAN GILLETTE levels vital to computing. Two computers may have different codes for letters, different methods of structuring stored information and messages, different access procedures, and, generally, different protocols. Unless the whole telematics infrastructure is built to a single set of standards and protocols, an interface message processor (IMP) may be needed between a terminal and the network. Aggressive development of personal computers (PCs) in the last few years has made the notion of a dumb terminal almost obsolete. PCs with startlingly great capabilities are now advertised for a few hundred dollars- prices lower than those of dumb terminals only a few years ago. Because software for logical interaction over a telematics network can now be lodged in a relatively inexpensive smart terminal, the need for IMPs should decline. Distributed Processing The first digital computers sat in solemn majesty in air-conditioned space, and users, like acolytes, traveled to them. Comparisons with ancient oracles were obvious. With the advent of telematics, not only could users communicate instead of traveling, but computers could be joined for work in concert on a common problem. The hierarchy of computers that support a large bank's statewide distribution of automatic teller machines is an example of such distributed processing. The concept of geographical distribution is powerful, but exploitation of the idea to its fullest extent is difficult. One thorny problem, for ex- ample, is that of storing widely needed data. Should storage be in one place with relatively expensive communications to each computer, or should each computer store part of the data and use communications for limited access and updating? If the latter method is used, what is the logic -of distribution and the plan for changes? Parallel processing, a common term in contemporary computer design, refers more directly to the use of multiple processing units within a com- puter than to the use of distributed processing. Development of strategies and programming tools for efficient parallel processing is a formidable task expected to lead to new levels of processing speed. Such a task is distinct, however, from the effort that arranges the geographical distri- bution of processing that characterizes telematics. Standards and Protocols Interconnection standards are necessary in any physical system: rail separation in railroads, voltage level and frequency precision in power

OCR for page 233
COMBINING COMMUNICATIONS AND COMPUTERS 239 distribution, maximum pressure and pipe thread shape in water systems, frequencies of the tones in push-button telephone systems, number of lines in a television picture, and on and on. Circuit-switched telephone networks need another layer of standards to define the process by which a connection is requested and completed. After a voice channel connection is made, however, a voice circuit is free of protocols except for limitations on bandwidth. Any language can be used, speech can be fast or slow, and two people can even talk simulta- neously, if incomprehensibly. In data communication, switched or not, additional layers of standards are needed for compatibility. Such standards cover data rates, the repre- sentation of letters and numbers, the structure in a block of bits, and the length of a packet. All such questions must be answered unambiguously before a digitally encoded message can be transmitted. In telematics, yet another level of standards is needed to achieve software compatibility among computers and between computers and terminals that may have access to several different kinds of computers. Some aspects of standards and protocols are susceptible to analysis: choice of analog-to-digital encoding for reproducibility, for instance. Other choices, such as block length in packet switching, are more arbitrary. If one company manufactures all the parts of a system or can completely specify a portion for an outside manufacturer, standards may be set by that company, and any arbitrary choices will be determined by fiat. Three examples are AT&T's standards for its nationwide telephone-to-telephone public switched network before the Federal Communications Commission (FCC) permitted customer-owned terminals and competitive long-distance service; IBM's System Network Architecture (SNA) for extensive tele- matics systems; and Xerox's Ethernet standard for local area networks. Occasionally, a private standard becomes a de facto industrywide standard, either because the initiating entity dominates the industry, as AT&T did for its telephone network plan, or because it is more convenient to adopt an existing standard than to develop a new one, as some other suppliers found in the case of Xerox's Ethernet. When a system is not the sole responsibility of a single supplier, a forum for agreement among the many suppliers must be established. In the United States, standards may be set by professional societies (for example, the Institute of Electrical and Electronics Engineers LIEEEl) industry associations (for example, the Electronic Industry Association), government agencies, or independent groups. The American National Standards Institute (ANSI) has established rules for the structure of stan- dards groups and the processes by which they develop and publish stan- dards.

OCR for page 233
240 DEAN GILLETTE The International Telecommunications Union (ITU), originally estab- lished as the International Telegraph Union and now headquartered in Geneva as part of the United Nations, has been the forum for setting transborder standards for nearly 120 years. "Recommendations" for data communications are made by the Consulting Committee for International Telephone and Telegraph (CCITT). The CCITT "X.25" recommended protocol for packet switching, for example, is now almost universally used. The International Standards Organization, which is also based in Ge- neva, provides another forum. Its recently published open systems inter- connection (OSI) reference model for computer-to-computer communications quickly became the foundation of telematics protocols. The OSI model is an architecture that distinguishes seven layers of standards and protocols so that the task of setting standards is less likely to drift into irrelevancies. The lowest three levels are primarily concerned with data communications, and the upper three with the software issues unique to telematics. The intermediate fourth level involves data processing that helps control com- munications.5 ARRANGEMENT OF TELEMATICS SYSTEMS Topologies and Data Flows Data transmission channels have directionality. For example, a 1.5- Mb/s digital channel carried on a pair of wires with repeaters spaced at ~-mi intervals transmits data in only one direction; for two-way com- munication a second pair of wires is needed. A one-way data channel is referred to as "simplex. " A two-way channel is "full duplex," whether it uses a single medium, as in telephony, or a coordinated pair of media, such as the two simplex channels in a 1.5- Mb/s digital trunk. "Half duplex" uses a single medium alternately for transmission and reception, as in "press-to-talk" citizens' band radio . . communications. The basic topologies of telematics systems are the star, bus, tree, ring, and compound arrangements shown in Figure 10-1. In the star network, all terminals, indicated by T's in boxes, are connected to a central point, and the connections must be duplex. In the bus topology, all terminals are connected to a single medium, which must be duplex; the ring is like the bus except that it uses simplex transmission. A tree might be used for one-way distribution from a single node, as in cable television; a return channel requires duplex transmission. In the star, bus, tree, and ring topologies, only one path exists from one terminal to another. In the

OCR for page 233
COMBINING COMMUNICATIONS AND COMPUTERS [it STAR ~ _7 3--~:, RING BUS lo l TREE FIGURE 1 0- 1 241 ~- E~ COMPOUND Local area network topologies. compound topology, any of several paths can exist from one terminal to another, and alternate routing is possible to avoid an overload or a damaged facility. Either simplex or full duplex channels are used. In the star and compound topologies, switching functions at the nodes control traffic flow. They use either circuit or packet switches, but they assign channels and times of transmission uniquely according to internal protocols and the addresses presented. The compound arrangement is characteristic of the public switched telephone network, which uses full channels, and nationwide packet-switched networks, which use simplex channels. The star is characteristic of a switching office serving local telephones or a telematics system in which point-of-sale terminals are polled in turn to ask if they have a transaction to carry out. In the bus and ring topologies, the control of data flow is distributed. Each terminal may be assigned a unique time slot or frequency for trans- mission, and each "listens" to all slots and frequencies for messages addressed to it. These protocols are referred to, respectively, as time division multiple access (TDMA) and frequency division multiplexing

OCR for page 233
242 DEAN GILLETTE (FDM). An alternative is to provide a single channel accessible by all terminals and establish a protocol to avoid simultaneous transmission. One such protocol is token passing, in which transmission is permitted only by the terminal that has the token or password. If that terminal has no message to transmit, or after it transmits a specified block length of data, it passes the token to the next terminal in line. Another popular protocol is carrier sense multiple access with collision detection (CSMA/CD), under which all terminals listen for a signal generated by the channel when a message is being transmitted. If a terminal has a message to send and the carrier signal is absent, it transmits. If two terminals transmit simulta- neously, each can detect the other's signal and that a collision has occurred. In that event, both terminals stop transmitting, wait a randomly determined period of time, then transmit again if the channel is clear. Networks Local area networks (LANs) are, as the name implies, data networks that serve a small area, such as a university campus or an industrial park. A star topology entered through telephone lines in a local private branch exchange may be used if data rates are low and the efficiency of packet switching is not needed. More frequently, LANs use a bus or ring topology with high-speed channels in coaxial cable or optical fibers. One example is Ethernet, which sends data at 10 Mb/s on a coaxial cable in a bus configuration and uses the CSMA/CD protocol. Nationwide networks provide communications among LANs and among widely dispersed individual users and computers. Some nationwide net- works are private, such as those in military applications, and are built to a single set of protocols. Other networks are public and must accommodate users with a variety of equipment and connecting LANs with a variety of protocols. Interface message processors may be needed to connect ter- minals to the network, and "gateways" may be needed to connect one network to another. Both are responsible for format and protocol con- version. One example of a nationwide, indeed worldwide, network is BITNET, which was started in 1981 by Yale University and the City University of New York with support from IBM Corporation. It operates at 9.6 kb/s and is a store-and-forward message network. Other public networks are the Advanced Research Projects Agency Network (ARPANET) and GTE Corporation's TELENET.6

OCR for page 233
COMBINING COMMUNICATIONS AND COMPUTERS SOCIAL ISSUES 243 The development and application of the computer has been accompanied by social concerns about invasions of privacy, job displacement, the di- chotomy between information-rich and information-poor populations in the United States, and a new form of Third World economic lag. The introduction of telematics has not only exacerbated these problems but has brought new ones, the most prominent of which are remote computer intrusion and a slowdown in the innovation process. Computer Intrusion Tapping a telephone line to listen to people talking has been a federal offense for years, and in the fall of 1986 it became a federal offense to tap a data stream. Although there are state laws, as in California, that prohibit the unauthorized use of a computer, there are also federal sanctions for an electronic intruder in, for example, Arizona who readjusts financial accounts in a California bank. Computer intrusion for the discovery of personal facts (invasion of privacy), for financial gain (theft), or merely for the "pleasure" of prying or to prove that entrance is possible ("hack- ing") has become prevalent enough to be a plot theme in popular movies. Fortunately, owners of data bases have become aware of their vulnerability if telephone line access is possible and are developing security measures. The ultimate security, of course, is to use closely guarded, private, ded- icated communication channels or not to use telematics at all but to return to the central, noncommunicating computer, exchanging efficiency for security. The Innovation Process Before the advent of telematics the markets of the computer and tele- communications industries did not overlap. AT&T was barred from any business except communications by the 1956 consent decree that con- cluded the federal antitrust suit begun in 1949 against the company. Com- puter companies, on the other hand, were busy enough exploiting a rapidly advancing data processing technology. As telematics emerged as the suc- cessor to stand-alone computing, however, the distinction between com- munications and computing began to blur. As this convergence of technologies was occurring, the nation's reg- ulatory bodies were becoming disenchanted with regulation itself and were seeking ways to introduce competition as a preferred mechanism to protect

OCR for page 233
244 DEAN GILLETTE consumers. One consequent action was to force AT&T to divest itself of its local telephone operations in return for relief from the provisions of the 1956 decreed The expected result was "fairer" competition in intercity telecommunications and increased competition in computing. With the removal of AT&T as the proprietor of the bulk of the national telecommunications network, two changes in the evolution of telematics have occurred: AT&T can no longer set de facto protocols and standards for the nation, and the regulatory pressure to keep depreciation rates low has evaporated. The result of the first change is that revision of standards and protocols will be slowed by the need to have competitors agree, although, as discussed earlier in this chapter, there are already forums for such deliberations. The possibility of more rapid depreciation may be more subtle. When AT&T had its monopoly under regulation, it designed, built, and installed equipment for long lifetimes and low maintenance costs. Two hours of downtime in 40 years was an objective in the development of switching systems. Annual operating expenses were low, consonant with a regulatory desire, but the introduction of new services was also slow. With regulatory restraint gone, less will be spent on long-lived equipment, and new services will be introduced more rapidly. Shorter lifetimes, however, are equivalent to more frequent repairs. Together with the elimination of the Bell System as the end-to-end service supplier, these factors have made it more im- portant for telematics users to develop the competence to select com- munication options from the variety available in the market. The management task is not new; it is similar to that needed in selecting vehicles for car pools and computers for corporate support. PROSPECTS FOR THE FUTURE Components Solid-state electronics technology is fundamental to all physical aspects of telematics, including computers, switching systems, and transmission systems. The current vigorous and fruitful research and development in this area indicate that there will be at least another decade of improvements to capability and reductions in cost.8 Transmission systems of greater capacity and lower cost will continue to be developed as the costs of integrated circuits decrease and manufac- turing experience is gained, particularly in optical fibers. There will prob- ably be a pause in additions to the armada of U.S. communications satellites as the space shuttle program is restructured and as new unmanned launch

OCR for page 233
COMBINING COMMUNICATIONS AND COMPUTERS 245 vehicles are developed, but the deployment of optical fiber systems should bridge most gaps in the interim.9 As digital switches replace space division switches in the telephone network, the time needed to carry out circuit switching will decrease. Further improvement will occur as central control of interoffice switching (CCIS) is extended. CCIS is a scheme in which the addresses, or numbers, of the calling and called parties are sent to a central computer connected by data links to all switches in a network. The computer finds a route for the desired connection and orders all the necessary switches to close simultaneously. Connection times can be in tens of milliseconds. Both switching systems and computers will benefit from improvements in integrated circuits. Although development of even faster and more capacious computers will increase the demand for improved data com- munication networks and more sophisticated gateways and IMPs, the technologies seem to be available to meet that demand. Thus, substantive progress is evident in most technologies fundamental to advances in the physical aspects of telematics. Yet improvement in terminal apparatus that is now expensive or cumbersome still awaits new discoveries or technical approaches. True end-to-end electronic mail will not become a reality until a combined typewriter and hard-copy printer is available for under $100, and video display units will continue to occupy a substantial portion of a desk until a flat screen display as flexible and inexpensive as a cathode ray tube is created. The current diversity of protocols and standards causes difficulties and extra expense in arranging telematics systems. Exploitation of the seven- level OSI protocol structure should help avoid incompatibilities. Some of this is already happening. Digital Equipment Corporation, for example, intends to modify its DECnet protocols to conform to OSI, and both the Boeing Company and General Motors Corporation have adopted OSI for their private networks. Facilities Arrangements Cost reductions in transmission may come more swiftly to some users through the rearrangement of existing facilities than through the devel- opment of new ones. For example, buying and installing a ground station for transmitting and receiving satellite communications to serve an office complex may be less expensive than leasing the equivalent capacity from telephone companies. This method is justifiably called the bypass method. Its use is a political issue because regulators are concerned that the loss of the profits from overpriced business telecommunication services will force local telephone companies to raise the price of residential services

OCR for page 233
246 DEAN GILLETTE to compensate for losses due to underpricing in that market. Regulators have taken little corrective action, however, and businesses are increasing their use of bypass facilities for many long-distance communication ser- vices, including telematics. U.S. regulators have limited the options of some telematics system designers for example, through FCC decisions that prohibit AT&T from making any protocol conversions and from using some level 4 protocols in structuring networks for the services it offers others. In principle, however, another entity could provide the IMPs and gateways and then lease digital circuits from AT&T to provide a complete service after negotiating with AT&T for interconnection protocols. AT&T, of course, is trying to convince the FCC that the restrictions should be removed. Social Issues The social issues accompanying the introduction of telematics are being addressed by appropriate legislative, institutional, and business groups. Opportunities for damage by computer intrusion should decrease as new laws are passed and as data base managers introduce access controls and tighten security. The market issues are more difficult to assess, however, because the telematics industry structure was so altered by the breakup of AT&T. We will never know how rapidly innovation would have occurred in the old framework and can only hope that delays in the revision of standards will be outpaced by advances in technical developments. Per- haps, too, the improvements in reliability that are accompanying the new- est in technology will help keep service levels up as we enter an era of deliberate obsolescence in telematics equipment. Continuing Evolution The agent for continued evolution of telematics infrastructures is tech- nology, particularly solid-state electronics and software. Strong research programs in both areas can be found in private industry as well as in universities. The federal government both supports research and encour- ages private industry to create new services. And private industry is eager to accept the opportunity to capitalize on a growing need. Nor is the United States alone in providing new telematics services: foreign entities either provide or support the development of telematics systems. Nippon Telephone and Telegraph's CS-1 and CS-2 communication satellites re- spectively provide transparent communication and teleservices using the OSI protocol structure. The European Research Community, the United Kingdom, Ireland, Australia, and Canada are also continuing aggressive development of telematics systems.

OCR for page 233
COMBINING COMMUNICATIONS AND COMPUTERS 247 Moreover, in cooperation with the CCITT, various nations are devel- oping plans for an integrated services digital network (ISDN) that even- tually will provide voice, data, and image transmission services over a single network incorporating both packet and circuit switching with CCIS control. The deployment of a 64-kb/s voice and data transmission capa- bility is expected in the United States in the late 1980s. Overall, it appears that telematics infrastructures can serve today's needs well; it also appears that foundations to meet our future needs are in hand or are being devel- oped. NOTES 1. Two analyses that both established the 50 percent figure but were carried out in entirely separate ways are in Porat (1976) and Abler (1977). In the decade since this work, U.S. information industries have continued to grow while jobs in manufacturing and agri- culture have declined; thus, the former has become even more critical to the U.S. economy. 2. This' reference is a report to the president of France on the importance of telematics to the French economy. The translation (Nora and Mine, 1980) includes a foreword by Daniel Bell. 3. Multiplexing is a technique for reducing the cost of transmission by using a single medium to communicate several messages simultaneously. In frequency division mul- tiplexing (FDM), each message is impressed on a different carrier frequency at the transmitter and separated at the receiver by a tuner that responds only to that frequency. One application is found in contemporary cable television systems, which use FDM to carry scores of TV programs to a residential channel selector. FDM is the oldest of the multiplexing methods. In fact, Alexander Graham Bell tried to send several telegraph signals over a single pair of wires by having each turn a unique frequency on or off to send the elements of a message. His experimental apparatus for such a "harmonic telegraph" was actually the one that carried the famous first telephone message. His knowledge of the science of speech led him to understand the significance of the result, even though he was working on another problem. If each of several messages is encoded as a string of ones and zeros, as in a data signal, time division multiplexing (TDM) can be used. With TDM a single high-speed transmission medium carries sections of each of several lower speed channels alternately in time. A T1 channel, for example, carries 8 bits from channel 1, then ~ bits from channel 2, and so on for 24 channels; the T1 channel then starts again with the next 8 bits from channel 1. As the natural complement of digital transmission, TDM has become the dominant technique in new communications systems. 4. The earliest of the switching technologies were electromechanical systems, which used relays for both the control function and the network connection function. Although data signals at a few hundred bits per second will pass easily through a relay connection network, the logic in relay control systems is simple and inflexible and not adaptable to the more exotic demands of some data communications. Electronic switching was introduced in the early 1960s when relays were used for connections, but the control function was a special-purpose, stored-program computer that could accommodate com- plex logic and software changes. Both electromechanical and electronic switches are referred to as space division switches because the input and output channels for both

OCR for page 233
248 DEAL! GILLETTE consist of discrete wire pairs with which the switches make physical, electrically con- tinuous connections. Either switch makes connections in tenths of a second. In the mid-1970s, digital switches were introduced. The control function in such switches is, again, a computer controlled by a stored program, but connections are made in a totally different fashion. Input and output channels are time slots in a time division carrier system. Connections are established by assigning the block of bits in a time slot . . In an incoming digital channel to a specified time slot in an outgoing digital channel, perhaps after a delay of a few hundred microseconds. The process is referred to as time slot interchange or time division switching and is the obvious complement to time division multiplexing or transmission. New connections can be established in milli- seconds. If incoming signals are on loops or analog carrier systems, they are converted to a digital format before presentation to the digital switch. If outgoing trunks are analog, a digital-to-analog converter is introduced between the switch and the trunk. Such conversion may seem cumbersome, but the complement must be accomplished if the incoming and outgoing signals to a space division switch are in a digital format. 5. Voelcker (1986) gives an excellent description of the OSI seven-level structure. 6. The periodical IEEE Communications is a good source of information on data networks. A new IEEE publication, Network, which first appeared in January 1987, is an excellent source for the newest in data communications. 7. Shooshan (1984) provides excellent essays on the economic, policy, and regulatory issues leading to divestiture by individuals responsible for many of the decisions. 8. Several relevant technological projections are given by Mayo (1985). 9. The potential for reduction in transmission cost is thoughtfully described by Lucky (1985). REFERENCES Abler, R. 1977. The telephone and the evolution of the American metropolitan system. P.1 in The Social Impact of the Telephone, I. De S. Pool, ed. Cambridge, Mass.: MIT Press. Lucky, R. W. 1985. Telecommunications research and development: A look at the next twenty years. Paper presented at the Fifth Convocation of Engineering Academies, London, June 10-15, 1985. Also in The Bridge 15(3):2-6. Mayo, J. S. 1985. The evolution of information technologies. Pp. 7-33 in Information Technologies and Social Transformation, B. R. Guile, ed. Washington, D.C.: National Academy Press. Nora, S., and A. Minc. 1980. The Computerization of Society: A Report to the President of France. Cambridge, Mass.: MIT Press. Porat, M. U. 1976. The Information Economy. Palo Alto, Calif.: Stanford University; and Washington, D.C.: U.S. Government Printing Office. Shooshan, H. M. III. 1984. Disconnecting Bell The Impact of the AT&T Divestiture. New York: Pergamon Press. Voelcker, J. 1986. Helping computers communicate. IEEE Spectrum 23(3):61-70.