Click for next page ( 8


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 7
The Evolution of Information Technologies JOHN S. MAYO THE INTERACTION OF TECHNOLOGY AND SOCIETY Humans were given capable and inquisitive minds, so they endlessly seek better ways of doing things. This dnve, coupled with an innate curiosity and a strong drive to unlock the secrets of nature, has created a steady stream of technical innovations over the ages. These innovative efforts have focused on the means for survival, comfort, and accumulation of wealtl~with the hierarchy of needs extending from physical basics of existence to higher-level wants associated with self-actualization. A principal thrust of innovation today continues toward technological advances that enhance the productivity of labor and free humans of tasks done more economically by machines. An insatiable appetite for convenience, comfort, and entertainment products and services, as well as for means to overcome natural barriers like geography and travel time, creates a constant pull on technology. The pull is especially strong in areas relating to the quality of life, and there have been many technical innovations to meet that need. But the opportunities are far from exhausted. Among society's newest demands on technology is for the means to handle the vast amount of information generated by modern life. This information explosion stems from sophisticated business practices, new residential services, substantially increased record keeping through extensive data bases, and the globalization of our advanced society. The information technologies have evolved over many years to assist a growing portion of the work force devoted to the generation, 7

OCR for page 7
8 JOHN S. AlAYO processing, transmission, storage, retrieval, and general use of infor- mation. Bureaucracies generated dunug the major wars and the rapid growth of social services in recent decades have helped increase the number of infonnation workers in the U.S. work force, producing a permanent change in our way of life. Stimulated by these and other spurts of rapid growth, the percentage of information workers in the U.S. work force has grown from about 10 percent in 1900 to about 30 percent in 1940 to about 50 percent in 1970. Since 1970 the fraction has held at roughly 50 percent, probably as a result of the new electronic infonnation technologies that augment human efforts. The computer, along with telecommunications, is making today's infor- mation work force more efficient, much as the engine raised produc- tivity during the industrial revolution. In both cases, society's thirst for technology to reduce labor was met in striking ways by a wide range of innovations of varying impacts. This thirst for technology creates a steady pull on innovation. In addition, the technologies themselves provide a push. From the families of all technologically feasible innovations of all time has come an almost endless reservoir of potential technology. However, between society's pull and the push of technology are two powerful gates, as shown in Figure 1. The technology available to society at any particular time is only that which can flow past the technology gate, which is Social TechnOlO9y _ Gate Gate _ Pull of society - \ mush of technology Survival \ ~ _ | AII feasible innovations Comfort 1. . _ | Limits of technology Quality of life / ~ _ \ Complexity many ~ \` t Economics R&D prowess Common good R&D management Public receptivity Embedded base Regulation and legislation Natural sequencing Standards FIGURE 1 The flow of innovations into society. SOURCE: AT&T Bell Laboratones.

OCR for page 7
THE EVOLUTION OF INFO~ATION TECHNOLOGIES 9 operated by a series of strong forces. Among them are R&D prowess' characteristics of the embedded technology base, natural sequencing constraints, and perceived standards limitations: . The force of R&D prowess is the sum of all the contributions of individual laboratones. The prowess of an R&D laboratory is limited by the skills of its scientists and engineers and by the capabilities of its support environ- ment, including both financial and physical resources. Laboratory prowess also clearly depends on the wisdom and judgment of the R&D management team. The embedded base of technology, such as existing systems or production facilities that represent a large investment, impacts the characteristics of the R&D laboratories and the factories that make their products. It often leads the R&D laboratories to seek innovations that have Meat synergy with the embedded base, which can be either a curse or a blessing. On the one hand, this force can limit the introduction of new technology as well as discourage breakthroughs in totally new directions. On the other hand, it brings focus and resources. With good systems engineering, older technologies can be phased out, current ones upgraded, and entirely new ones introduced, all in ways that are synergistic with the embedded base. Natural sequencing simply means that the invention of the integrated circuit, for example, would have been unlikely before the invention and development of the transistor. The imposition of standards prior to innovation can narrow the technology gate by forcing R&D laboratories to focus on innovations that meet preconceived standards, but which may in the long Ban not be the best innovations at all. Innovations that pass the technology gate must also pass the social gate. The forces that operate the social gate include economics, the common good, public receptivity, and regulation and legislation: . . The economic force depends not only on the marketplace, but also on the national economic structure. Currently, we see a significant difference in the way economic forces affect innovation in the United States as compared to Japan. Closely related to economics, but not always in concert with it, is the force that makes technology serve the common good. Society will even- tually, for the most part, either ignore or legislate against technology that does not serve the common good. The issue of public receptivity is related to that of the common good. In the United States, the public defeated the supersonic transport and appears to have nuclear power on its deathbed; however, it still remains to be seen whether such innovations do not indeed serve the common good. Regulation and legislation have been and remain powerful forces at the social gate forces very active in throttling technology on behalf of society.

OCR for page 7
10 . JOHJ!; S. AtA YO Much social good has come from such actions, but not without frequent adverse impact in the long run. The forces operating at the social gate are extremely powerful in selecting the innovations that actually succeed. They serve as a "tollgate" in the gap between the push of technology and the pull of society. The Bating forces will be further examined, following the discussion of the information technologies pushing at the gate. Technologies that survive both gates have primarily three types of impacts in the society they enter, depending heavily upon their character. First, of greatest impact are the "killer" technologies such as the engine, which replaced the horse, and the transistor, which replaced the vacuum tube. Their impact, of course, extends far beyond these major replacements, to opening whole new fields of opportunity. Included among these are opportunities to satisfy previously unknown or unrecognized societal needs and wants, often of an increasingly sophisticated nature. Second in impact are the "new domain" technologies. Although they do not replace earlier technologies, they do open up entirely new areas of opportunity. An example of the new domain technologies is automatic speech recognition and synthesis, a rapidly developing technology that win eventually allow inanimate objects such as cars and appliances to speak and listen much as humans do. Third, there are the "niche" technologies, which play a very important role in meeting society's needs. When they first appear, however, they are often mistaken for killer technologies. For example, when broadcast television became feasible, many expected it to kill newspapers, radio, and movies. Instead, it found its own niche and satisfied a thirst in society not previously meter perhaps even recognized. INFORMATION TECHNOLOGIES AND THEIR LIMITS The growth in information jobs cited above is but one major indicator of the rapid transition of our society to an information base. Another major indicator is the rapid growth in information technologies. George Stibitz's invention of the first digital computer, the achievement of universal telephone service, and the invention of solid-state electronics paved the way to the Information Age. The technology of the Infor- mation Age is digital. The information is represented as digits, which are generated, processed, transported, stored, recovered, and displayed in order to do useful things for humans. The key technologies for

OCR for page 7
THE EVOLUTION OF INFORMATION TECH!IOLOGIES 11 manipulating digits are integrated circuits, computing technology, software, and photon~cs, as discussed below. Integrated Circuits The most powerful force of technology today is the expanding capabilities of silicon integrated circuits. A tiny chip of silicon can contain an electronic circuit consisting of hundreds of thousands of transistors and all the necessary interconnecting conductorsand its cost is only a few dollars. The circuitry on that chip is equivalent to about 10 years of work by a person soldering discrete components onto printed wiring boards. It is this tremendous unprovement in the economics of circuit assembly, coupled with similar improvements in the reliability of individual circuit functions, that accounts for the power of this technology. A common measure of progress in integrated circuit technology is the number of components that can be squeezed into a single-chip circuit. Figure 2 shows the exponential growth in components per chip over the past two decades, and a projection for the next decade. The number of components per chip of silicon is still increasing by a factor of 100 per decade. Today the limit is almost 1 million components on a chip; by 1990, it will be at least 5 million; and by the year 2000, between 10 and 100 million. 109 108 O7 COMPONENTS PER CHIP 1o6 105 104 103 1o2 10 YEAR PHYSICAL LIMIT X100 PER DECADE_ X1000 PER DECADE ~ 1 1 1960 1970 1980 1990 FIGURE 2 Changes in component density for silicon production, 1960 to 1990. SOURCE: AT&T Bell Laboratones.

OCR for page 7
12 JOHNS. MAYO The limits of integrated circuit technology as we know it today are determined basically by how big we can make a chip and how small we can make the line widths used to define a working transistor. These limits can easily be estimated by assuming the largest practical chip to be about 1,000 square millimeters, and the smallest transistor to be fabricated using O.1-micrometer (am) line widths (a length of about 400 silicon atoms). When reasonable space for electrical isolation and interconnections is allowed, the resulting limit is easily derived to be about 100 million components per chip. For such a chip, using known technology, electrical isolation and interconnections would consume approximately 90 percent of the chip area. The magic of the ever-expanding capabilities of integrated circuits will therefore be with us for at least another decade. Component reliability will continue to increase dramatically, and integrated circuit chips will perform more and more functions, ever faster and cheaper. This progress will make possible increasingly powerful, reliable, lower- cost digital systems and much more flexible approaches to systems design. Integrated circuit progress is making it possible to have digital systems everywhere, be they for computing, robotic control, office automation, or telecommunications. Clearly, this force will continue to be a major spur to further progress in the information technologies. Computing Technology Computing technology is a major beneficiary of the power of integrated circuits. Figure 3 shows the past trends in computer processing power and forecasts the future. Processing power is ex- pressed in millions of instructions per second (MIPS), and each data point in the figure represents a specific computer introduced into the marketplace. Most notable is the rapid progress in processing power of the single-chip computer or microcomputer, which has been doubling each year. Extrapolation of the trend of the last decade forecasts that microcomputers will have processing power comparable to that of minicomputers and large, general-puIpose mainframe computers (max- icomputers) by the early l990s. Because the computer is the "engine" of the Information Age, having the power of today's largest computers on a chip or even a few chips of silicon to go in automobiles, appliances, toys, offices, factories, and homes is a tremendous driving force. The trend curve of components per chip in Figure 2 showed that by 1990 integrated circuits will be within a factor of 10 of the physical limit of the technology known today. That implies the maxicomputers, minicomputers, and single-chip microcomputers as we now know them

OCR for page 7
THE EVOLUTION OF INFORAlAT10N TECHNOLOGIES 10 2 10 PROCESSING POWER (IVIIPS) 1O-1 lo-2 1 1 IVIAXIS MINIS MICROS 1970 1980 YEAR FIGURE 3 Trends in processing power of computers, 1970 to 1990. SOURCE: AT&T Bell Laboratones. 13 1990 have an ultimate limit of a few lOs of MIPS. However, the limit of chip processing power is not a limit to the processing power of computing. For computing is rapidly moving toward new architectures involving multiplicities of processing elements such as single-chip computers. Multiplicities of such chips, however, are inherently more expensivelargely because of the high costs and difficulty of inter- connecting and programming them to function as single systems. For automobiles, appliances, tools, home computers, and the like, we can assume that most will operate with an ultimate computing power of the order of 10 MIPS or less per computer. But an automobile, for example, may eventually have a dozen or more computers. Software Software is vital not only to the operation of Information Age systems, but also to their interlinking with each other, with data bases, and with people. The demands for software are growing explosively for tailoring systems to customers' needs, for making them reliable, and for making them easier to use, or "friendlier.'' These demands, in turn, are leading to increasingly complex software, ironically, to achieve user simplicity. Unfortunately, software is the "bottleneck" information technology. Currently, it is generated principally by people, and most enterprises generate more software by hiring more people a very difficult and costly approach. It still takes a programmer approximately one year to produce a few thousand lines of code. In telecommunications, a

OCR for page 7
14 JOHN S. MAYO large electronic toll switching machine uses more than 2 million lines, a local electronic switcher more than 1 million lines, and System 85, a Private Branch Exchange (PBX), almost 2 million lines. Industry has reamed in the last few years how to manage big software systems, developing them to meet cost, time, and performance objectives. But it desperately needs an improvement in programming productivity to sustain both the growth in complexity and the increasing demand for software systems. The rate of progress in improving programming productivity remains extremely low. Figure 4 compares AT&T Bell Laboratones' produc- tivity growth for producing software with that for design of silicon integrated circuit chips and circuit packs. The bad news is common throughout the industry. Software productivity is improving very slowly. The good news is the increasing productivity of the electronics designers, and that is largely a software success story. In fact, without computer-aided design, much of the progress in todays most important technology would be impossibl~and tomo~Tow's tasks, hopeless. Even though chip complexity has increased 100-fold over the past decade, computer-automated tools permitted us to use the same design effort, as well as to significantly improve our ability to get error-free chip designs. There will eventually be dramatic improvements in programming productivity. They will come from continued improvements in com- puter aids for software design, leading eventually to automatic gen- 7 PRODUCTIVITY RELATIVE TO 1975 4 3 64~Icol:: 5 ~ ~ ~: UIT PACKS 24~0FTWARE ~3 '75 '76 '77 '78 '79 '80 '8 1 '82 '83 '84 YEAR OF DESIGN START FIGURE 4 Hardware and software productivity, 1975 to 1984. SOURCE: AT&T Bell Laboratones.

OCR for page 7
THE EVOLUTION OF INFORMATION TECHNOLOGIES 15 crayon of applications programs. Such a breakthrough in software productivity will first require development of durable and detailed technical standards, new methodologies for requirements generation, and large software design programs for structuring, generating, and testing code. Progress may first come through greater development of reusable software components with standardized interfaces. Even those standards do not exist and will be difficult to achieve. Early circuit designers rapidly solved that problem for hardware components. Clearly, it can also be done for software components. There is an argument, based on the analogy of motors, that today's software problems are transitory. When motors were new, users "hooked them up" to genders, saws, and numerous other elements to create functional tools. But as that technology matured, users' needs were met by functioning systems that contained motors~rills, grind- ers, saws, washing machines, dishwashers, cars, toothbrushes, toys an endless list. The analogy suggests that sooner or later a wide spectrum of software systems will be available so that most users will be able to buy functional Information Age products that perform the needed tasks. These products will just happen to contain software much as dishwashers and refrigerators, for example, just happen to contain motors. The user could not care less, so long as the dishes are clean and the food cold. Such functional software-based products are rapidly emerging in the marketplace, but the trend has just begun. Photonics Photonics is the key Information Age technology for transmitting large amounts of digital information. There are two key innovations: the laser and ultrapure glass fiber. Combined, they provide a trans- mission capability that far exceeds that of copper wire and radio to meet the most stringent needs of the Information Age. Photon~cs technology has progressed rapidly. In about a decade the technology has achieved some difficult technical milestones: Developing high-purity, ultratransparent, and high-strength glass fibers; Constructing long-life lasers that can operate at room temperature and at the appropriate wavelengths; Optimizing the mode of lightwave propagation in the fiber and shifting from multimode to single-mode fibers for many applications; - Determining and exploiting the unique wavelengths at which fiber trans- mission losses are the lowest; Developing means for wavelength multiplexing of multiple bit streams onto the same fiber; and

OCR for page 7
16 JOHNS. MAYO Developing single-frequency light sources which desensitize system per- formance to wavelength dispersion in the fiber. Where is innovation in photonics leading lightwave systems? The current technical frontier is in increasing bit rates. The basic trend continues toward higher communications capacity per fiber and greater distances between signal amplifiers or repeaters. For example, since AT&T introduced the first full-service commercial lightwave system in 1977, fiber capacity has increased almost 10-fold, from 672 calls per fiber to 6,048. The corresponding amplifier spacings have increased from about 8 kilometers to more than 30 kilometers. In laboratory experiments described recently, AT&T Bell Laboratones set a "dis- tance record" by transmitting 420 million bits per second over 125 miles without amplification. Also, 2 billion bits per second were transmitted over 80 miles using no amplification. That pulse rate can transmit the entire 30-volume Encyclopaedia Britannica in a few seconds. Underlying these accomplishments is healthy progress not only in glass fibers and lasers, but also in photodetectors and many other system components. The limit of loss in today's fibers is shown in Figure 5. The two wavelength bands at which the loss of practical fibers is both low and near the limit are in the regions around 1.3 and 1.5 ,um. Most new designs operate in these bands. Actual signal losses achieved in these bands are close enough to the theoretical losses that major resew breakthroughs in silicon fiber performance do not seem likely. On the other hand, advances in materials processing could lead to entirely new materials systems for fibers. Also, the power output of lasers will rise, and the sensitivity of signal detection subsystems will improve. So getting from today's capability of 125 miles without amplifiers to a few thousand miles without amplifiers may yet be feasible. The difference is not spectacular for domestic communications. But the capability of sending signals a few thousand miles without amplifiers is significant in globalizing the Information Age, for it would enable us to span the oceans with passive lightwave systems. What is the limit of lightwave technology as we know it today, and when win we reach that limit? Extrapolation of progress in the rate at which information can be sent through fibers and the distance it can travel without amplification, coupled with a little analysis, suggests the answers. Figure 6 shows that the product of rate in megabits per second (MBPS) and distance in kilometers (km) has been doubling yearly and this will probably continue for the rest of the decade, at least. Each data point in the figure represents the leading edge

OCR for page 7
THE EVOLUTION OF INFORAIATION TECHNOLOGIES 8.0 6.0 4.0 2.0 1.0 0.8 0.6 0.4 0.2 0.1 17 LOSS (dB/km) ' r MAXIMUM REPEATER SPACING {MILES) _ ~~ - ~ _ THEORETICAL ' l 1 1 - 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 WAVELENGTH {,wm} FIGURE 5 Lightwave communications technology. source: AT&T Bell Laboratones. 12 25 75 125 accomplishment for a single wavelength channelcontinually domi- nated by AT&T and the Japanese. Simple detection theory can be used to estimate the physical limit of today's lightwave technology. The estimate involves combining the theoretical loss and nonlinear behavior of glass fiber with an assumed maximum allowable laser power of approximately 1 watt and a minimum requirement of about 10 to 100 photons per pulse. This forecast suggests that the technology limits will permit the development of lightwave systems with each channel operating 10 to 100 times faster than today's best. Wavelength multiplexing will extend this limit by another factor of 10 to 100, giving an ultimate limit of about 109, or 1 billion, MBPS/km.

OCR for page 7
TlIE EVOLUTION OF INFO~ATION TECHNOLOGIES 23 Laser Materials Systems and Yields The materials systems for lasers are in a relatively primitive state today, resulting in rather low manufacturing yields. Improved materials systems and structures would greatly lower the cost of lasers and, in turn, accelerate the pace of advancement of lightwave technologies. Because molecular beam epitaxy permits control of materials down to the atomic levels, it offers exciting and expanding opportunities to custom-fabr~cate new materials that may lead to new device concepts, including lasers. TECHNOLOGY SELECTION AND THE PACE OF INNOVATION Now that we have a view of the key technologies, their limits, and potential new technologies, let us further examine the gating forces that determine how the winning technologies will be selected and the resulting pace of innovation. Marketplace Economics The dominant force pulling innovations through the technology and social gates today is the needs of the marketplace. For decades electronic technologies have been pulled into the marketplace as fast as humanly possible. Today's technology, however, is so rich that it can do more things than society might find useful. Increasingly, marketing resources are required to sort out innovations and potential innovations, to contain the scope of development, and to focus investment on the applications that will win in the marketplace. This growing force of marketing in the information technology arena is creating new and challenging relationships. Such a give-and-take relationship between marketers and technologists has long operated in low-technology fields such as soaps and toothpaste. Today, there is a similar, rapidly evolving relationship in fields of highest technology, especially in computers, software, and telecommunications. The pull of the marketplace rests largely on willingness to pay. To impact society significantly, an innovation must be of sufficient intrinsic value that users will not only pay the traditional manufacturing, sales, and related costs, but compensate for the high cost of development as well. Information Age products are very it&D-intensive. For example, R&D accounts for most of the cost of software, and the viability of software in the marketplace more strongly depends on sales volumes and copyright protection than do traditional manufactured products. Technology selection is strongly tied to cost trends. For example,

OCR for page 7
24 JOHN S. MAYO the proliferation of digital systems ties directly to the falling cost of a digital logic circuit, as shown by the curve in Figure 8. With each passing year, new digital systems become economically feasible be- cause their costs drop below what the user is willing to pay. The figure also shows that at 1 dollar per logic gate, the telecommunications industry could economically justify two digital systems: one called T1, for voice transmission on wire pairs, and a digital controller for the lESS local electronic switch. At 10 cents per logic gate, the 4ESS electronic switch, a large digital toll machine, and the Dimensions Private Branch Exchange became feasible. At 1 cent per logic gate, 5ESS Infonnation Age local digital switches became feasible, along with a wide variety of microprocessor-based "intelligent" telephones and terminals. Digital logic costs are now in the range of a tenth of a cent per gate and will be on the order of a millicent per gate by 1990. Each year as these logic costs continue to fall, the costs of a wide range of new Information Age products will drop below the threshold of user willingness to pay. The result is a mushrooming family of intelligent products- and the mushrooming phenomenon is not likely to slow down before the year 2000. The resulting economic climate will create a wide range of innovations, and economic forces will sort out the winners. A user's willingness to pay is not an absolute. So technology selection is strongly affected by public opinion and advertising. The case of 10 1 DOLLA RS . 1 PER CIRCUIT .0 1 .000 1 _ .00 1 .0000 1 T 1 DIGITAL \ TRANSMISSION - l 1, 1 1 1 1 1 1 4ESS DIGITAL TOLL SWITCH \ 1 ESS \ SESS ELECTRONIC ~ hi' I DIGITAL LOCAL SWITCH \LOCAL SWITCH DIMENSION \ , PBX \ \ 1950 1960 1970 1980 1 990 FIGURE 8 Distal integrated circuit costs, 1950 to 1990. SOURCE: AT&T Bell ~bo~tones. /

OCR for page 7
THE EVOLUTION OF INFORAlATION TECHI!iOLOGIES A_ videotape recorders versus the videodisk is an interesting example, even though they are not functionally equivalent. It seems that society has selected the videotape machine as the winning technology, possibly because of its recording capability. The totality of forces that produced that selection is very difficult to characterize, but I suggest that the technologist needs help if he or she expects to forecast the technology selection process. R&D Economics The force of economics is somewhat stronger at the social gate than it is at the technology gate, but even so it is a powerful force behind our R&D laboratories and even our university research centers. For at their limits the key technologies are extremely expensive to develop. The pace of innovation and, to a lesser extent, the direction of innovation are increasingly controlled by R&D economics. Clean rooms, feature patterning equipment, electron beam machines, and the like are terribly expensive at the micrometer level of capability and will become even more expensive as we move to submicrometer geometries. Small companies and universities are increasingly becom- ing followers rather than leaders. Today most universities cannot afford the requisite equipment, and even the wealthier ones are having to form special alliances with industry to raise the needed capital. The situation will ease somewhat as more universities find that complete fabrication facilities axe not essential for strong teaching programs or even specialized research. But the trend is still there: the leading edge R&D that produces the significant innovations that guide all of industry can be afforded by only a few institutions in our society. Government of course plays a large role in R&D economics. Of particular importance are the general economic climate of the nation, tax incentives, antitrust relief, and sharing of the output of the federal laboratories, as well as technology initiatives in the Department of Defense and the National Aeronautics and Space Administration. Each of these forces is managed on its own, but the sum of all the forces is not managed to speed the pace of technology and optimize the nation's technical position. R&D Prowess The prowess of our industrial, government, and university R&D laboratories remains a major factor in technology selection and the pace of innovation. Competence and motivation of individual scientists

OCR for page 7
26 JOHN S. MAYO and engineers are vital to R&D prowess, but management of those laboratories also is very important. A prerequisite for innovation in a particular area is the dedication of people and capital resources to that area, and creation of an atmosphere conducive to innovation. AT&T Bell Laboratories is widely acknowledged for having the winning combination, as well as for creating most of the innovations for the Information Age. Their inventions including solid-state devices, la- sers, and a wide range of telecommunications and information tech- nology spawned new industries that have gained significant innovative strength of their own. The power of these innovations to meet social needs is so great that they have forced change in some of society's major institutions. By opening up vast new frontiers of business opportunity and spawning numerous competitors, solid-state technology blended telecommuni- cations and data processing and led to the restructuring of those industries, just as the engine forced restructuring in the industrial revolution. Industry structure, in turn, is a strong force in technology selection and in pacing innovation. We are dealing with extremely powerful technologies whose forces are at best barely under our control. Continued U.S. industrial competitiveness is heavily dependent on the prowess of our R&D laboratories. Without a superior technical position, we could not be a force in technology selection or in the pace of innovation. The breakup of the Bell System has created a national challenge to ensure that the new environment has at least the R&D capabilities of the old. Certainly, AT&T is fully dedicated to continuing the innovative strengths of AT&T Bell Laboratories. However, the communications industry is much more fragmented than it was, and there are no means for managing it collectively. This is in sharp contrast to the major U.S. R&D competitor, Japan, which has numerous mechanisms for guiding its collective R&D efforts at the national level. Ironically, Japan has succeeded with its high technology by using methods, including statistical quality control, largely copied from the old Bell System. Ensuring the future prowess of our R&D laboratories will take more than blind faith that competition among our R&D enterprises will spur results that are more innovative and cover wider domains than the collectively managed enterprises abroad. Success will come from meeting two broad challenges posed by the new environment. First, our industries in general must continue to devote the needed resources to ensure steady streams of innovation from each R&D laboratory. And second, in this new environment, each of the R&O laboratories must continue to add its innovations promptly to the advancement of

OCR for page 7
THE EVOLUTION OF INCHOATION TECHNOLOGIES 27 science by continuing free and open dialogues within the technical communities. Regulation Few actions can throttle the flow of technology and slow the pace of innovation more than regulatory actions. In general, regulatory actions direct an entire industry or the most able within that industry. Actions are usually based on issues other than technology selection or pace, but the result may have enormous impact on technology. Regulating an entire industry always has the potential peril of handi- capping every one of the affected domestic companies and of benefiting every one of the foreign competitors. Regulators must acknowledge that the nation's major innovations will continue to come from its most able enterprises, and that through their action or inaction- they often have the power to throttle the outputs of these enterprises. This throttling of technology comes from action as well as inaction. For example, legislative action ruled AT&T out of the international satellite business and thus led to the disbanding of the pioneering Telstar satellite development team, an act that hardly accelerated the flow of satellite technology to the marketplace. Regulatory inaction delayed the availability of cellular radiotelephones to the public by a decade after AT&T had the technology available. At AT&T we are still experiencing a regulatory environment that restricts information flow within our business and favors competitors who have not contributed to the technology and have trivial R`&D capabilities. Such regulation places us in an environment that is insensitive to the funding of R&D that can malice a contribution not only to our business, but to the nation at large. As hard as we try to offset these handicaps, they remain a factor in our ability to maintain the flow of innovation. Technical Standards Technical standards are vital to the evolution of the Information Age. New products must not only be compatible with each other, but also with older products. And the standards must be firm enough to ensure compatibility, and prompt enough to ensure rapid introduction of new technology. Indeed, rapid progress seems to depend on prompt standards or no standards. Almost-at-hand standards encourage de- velopers to wait and discourage researchers working on alternative options.

OCR for page 7
28 JOHN S. MAYO In the past, key industry leaders developed the technology and brought it promptly to the marketplace. AT&T technology rapidly became the standard for the telecommunications industry, and IBM set powerful standards in the computing field. In today's new environ- ment, no single enterprise can pioneer the standards. For example, AT&T's UNIX operating system grew out of telecommunications, but it is rapidly becoming a worldwide computing standard. Fu~thennore, telecommunications has `' shrunk '' the globe, so stand- ards must increasingly be global. In working to set global standards, the technologist has encountered severe political issues, such as "one nation, one vote." Attention to the global marketplace has created an increasing tendency to set up technical standards before new products emerge. That approach has great impact on technology selection and is sure to delay the flow of new technology because of the substantial risk in introducing products in parallel with the standards deliberations. Unfortunately, these frustrations, which have plagued the international standards scene for years, are now characterizing the domestic scene. The political motivations in domestic standard setting are greatly enhanced in the new environment, and the ability of domestic standards bodies to handle the issues is yet to be proven. A major challenge is to try to separate the political and technical issues, with the hope that the technical issues can be resolved promptly. The Embedded Technology Base Three examples illustrate the force of the embedded base: silicon circuits, lightwave systems, and magnetic bubbles. The world's investment in silicon circuit fabrication facilities is estimated to be $15 billion. Most electronic products in the field are based on silicon, and the popular "direction" of the past decade has been, "If it can be done in silicon, it will." Until recently, almost 100 percent of the integrated circuit R&D was silicon based, partly due to the momentum of the embedded base, and partly due to the power of that technology. Lightwave is a counterexample. Although there was no embedded base, enormous R&D resources have been devoted to that technology. The marketplace views lightwave as a killer technology, and tends to favor it over the embedded base. Magnetic bubble technology was a potential killer that has been forced, at least for the present, to a relatively minor role. The power of the embedded base of magnetic disks and silicon random access memories drove cost reductions in these technologies as steep as the

OCR for page 7
THE EVOLUTION OF INFO~AT10N T~CHNOLOCIES 29 learning curve for bubbles. As a result, R&D in bubble technology diminished over the years and its learning curve slowed. So it will be a long time, if ever, before magnetic bubbles displace the embedded base. Josephson junction and other cryogenic devices appear to suffer the same fate, though they have never gotten as close to the competition as did magnetic bubbles. Both bubbles and cryogenics appear to be niche technologies at best. THE INFORMATION AGE Let me conclude by looking ahead. Considering the information technologies and their limits, pending new technologies, and forces controlling selection and pace, we can construct a pretty good picture of the technology of tomorrow. Whether this technology comes fast or slow, from our traditional R&D laboratories or from Japan or elsewhere, it is clearly leading us quickly into the Information Age. What does this mean to society? This section considers some of the expected changes when information technologies will assist the mind much as the industrial revolution's technology assists the muscle. Computers Everywhere The industrial revolution produced the now mature age of motors- which we take for granted. All of us are surrounded by motors, from the large ones in our cars and boats to the small ones scattered about our households. As the Information Age matures, we will increasingly be surrounded by computers, from large ones handling major home or business tasks to a multitude of microcomputers in our cars, appliances, toys, games, entertainment centers potentially everywhere. And we will think no more of computers lying idle in gadgets we hardly ever use, than we Worry "~out unused motors in our homes today. Computers are becoming so widespread and low cost that we can afford to take them for granted, too. Overcoming Geography The Information Age has another dimensio~bndging distance or geography. The interlinking of computers, people, machines, and data bases by the telecommunications network adds a new dimension of excitement. The Information Age will probably not offer electronic transport of matter, but it will do the next best thing. It will sense and transmit the remote environment to your presence via audio, video,

OCR for page 7
30 JOHNS. MAYO and data. Such benefits depend in large part on the increasingly digital nature of the electronic world, with digital connectivity becoming universally available. Independent of geography, then, machines can talk to machines, machines can talk to people and people to machines, andof course, as usual people can talk to people, but with expanded options such as video conferencing. It is increasingly feasible for people to conduct business from wherever they happen to be by accessing more and more capable telecommunications services and links back to their usual points of contact. One can even envision a "telepresence" using sensor-equipped robots to transmit perceptions back to the user. The user could "move" through scenes and even manipulate parts of the environment at long distance. New Services Information technology makes possible an expanding family of services such as financial transactions from home and office com- puters and entertainment, shopping, and accessing of data bases by video and telephone links. Electronic mail and the ability to leave telephone messages are developing, though slower than the technology allows. Increasingly economical equipment can control energy use in auto, home, and office; and it can monitor security and report alarms by the automatic dialing of emergency services. A wide range of direct- dialing, teleconferencing, and announcement services are available on the telecommunications network. The intelligent telecommunications network also makes it increas- ingly possible for users to have designated calls forwarded to another location, to have distinctive ringing identify calls from a specified set of telephone numbers, and to activate traces on nuisance calls. The network's intelligence also makes possible a wide variety of services such as the ability to dial one national number from anywhere in the country to be connected to the nearest emergency auto service, hospital, or any number of important services. Another potential innovation is to assign telephone numbers to people rather than telephones. By keeping the computer informed ofthe nearest telephone, you could designate calls to be routed automatically to youwherever you might be, at your office, home, car, or at some hotel while on a tnp. The same network intelligence is at the heart of cellular radio technology, which makes it potentially feasible for every car to have a telephone. This mobile phone capability is now spreading rapidly.

OCR for page 7
THE EVOLUTION OF INFORMATION TECHNOLOGIES 31 Of course it works just as well for personal portable phones as for car phones. Universal Information Service Early telephone pioneers wisely and quickly realized that achieving the full potential of the telephone depended on every phone being connectable with every other one. That realization is remarkable because it came when few people had phones, when phones were interconnected in small clusters to serve neighborhoods and businesses. And it came, despite the clear fact that any individual phone, over a lifetime of use, would actually need to connect to the tiniest fraction of the total telephone population. So it is for individual computers. The concept of universal information service recognizes that every computer, data base, or smart terminal must be connectable with every other, even though most of the possible connections will never actually be needed. Today, clusters of data networks are growing in selected neighborhoods and many businesses. Fortunately, the telephone network is able to interconnect these clusters. But, for a variety of reasons, total connectivity falls far short of what is needed for universal information service. Achieving this goal will be a great challenge for the information technology and service industries perhaps more so than universal telephone service was in the past. Video Data Bases High-speed digital transmissionusing the ever-increasing infor- mation-carrying capacity of lightwave systemscould make video communications and pictorial data bases as widespread as today's telephone service. And this prodigious capacity could also be used to carry high-speed data for processing by intelligent machines. These machines could search huge quantities of transmitted data, selecting and storing only the particular informationperhaps a news item, stock quote, message, or even a movie- that fits their users' needs or interests. Expert Systems The combination of increasingly powerful integrated circuits, com- puting technology, and software should enhance knowledge and allow increasingly expert software to take over some tasks done by expert

OCR for page 7
32 JOHN S. MAYO humans. An example is the expertise required to generate more accurate weather forecasts ranging from global forecasting to the kind of area forecasting that would permit us to manage crops better and to accurately predict the landfall of hurricanes. These expanded tech- nological capabilities will often be applied to create "expert" systems in services ranging from legal and financial advice to medical diagno- sis not necessarily to replace the doctor or lawyer, but to make such services available conveniently and inexpensively to large numbers of people. Expert systems exist today and perform specialized tasks in the telephone network. Engineering Perspective From an engineering perspective, the electronic equipment of the Information Age all looks much the same. It is digital systems made up mostly of integrated circuits mounted on printed wiring boards. The equipment is extremely compact in terms of number of gates per circuit board. Interface equipment is usually a keyboard, telephone lines, cathode ray tube, and/or liquid crystal display. Plasma panel and large liquid crystal displays will displace some of the cathode ray tubes, but not for full-motion video for some time. The equipment is increasingly reliable, and large systems will contain extensive diag- nostic subsystems for maintenance. Most reconfigurations and rear- rangements will be made via software rather than by manual operations. The equipment is increasingly lower in cost per function, so larger and larger systems will be of the throwaway type, much as low-function pocket calculators are today. The systems will increasingly be designed by machines through even more powerful computer aids to design. And, of course, the software content of the systems will grow, but hardware will also grow to ease software burdens. The challenge for engineers will shift toward the two ends of the spectrum of work: process development and computer-aided design systems to support design on the one end, and systems architecture and higher-level design on the other. The fast pace of technology, the exciting systems possibilities, and the expanding product opportunities put the Infor- mation Age designer in an enviable position. CONCLUSION The information technologies are far from exhausted, though the limits of the major technologies are well known. A number of new technologies are emerging, and progress should continue at a fast pace

OCR for page 7
THE EVOLUTION OF INFORMATION TECHNOLOGIES 33 for at least a decade or two. With the possible exception of integrated optics, during this time evolution is not likely to be dominated by new killer technologies. More likely, the rapid pace of current developments will continue to create ever more favorable economics, and extend the known technologies into new domains. The forces that control the pace of innovation and technology selection are not likely to change substantially unless the restructuring of the telephone industry produces unexpected results or overseas competition forces government action. Good innovations will continue to be rapidly pulled into the market- place. The resulting richness of high-quality, low-cost technology should help create a better society an Information Age with a host of new computing and telecommunications services to make life more pleasant, productive, and interesting. Comments ERNEST S. KUH Professor of Electrical Engineering University of California, Berkeley I would like to begin by proposing a simpleminded model of technology evolution for the mathematically inclined. Using the state-space analogy, which is familiar to most young electrical, mechanical, and aerospace engineers, we may represent the interaction of the four key elements of technology evolution that John Mayo defines: (1) technology base, (2) research and development, (3) sequencing, and (4) standards. In my proposed model, the state of the dynamic system corresponds to the technology base in Dr. Mayo's analysis; the input corresponds to R&D; the dynamics of the system correspond to sequencing; and finally, the set of constraints corresponds to standards. It might be possible then to use this analogy to introduce, for some technologies at least, a quantitative analysis of evolution through the technology gate. Models aside, the second part of Dr. Mayo's presentation gives a brief account of recent and prospective innovation in information processing technology. I would like to respond to that portion of the presentation with three comments. First, that which impresses me the most are advances in lightwave tech- nology. When I worked at Bell Laboratories 30 years ago, I was designing repeaters for submarine cable using vacuum tube technology. The progress made during the last 30 years in transmission is remarkable. Second, the technologies John Mayo did not discuss were such mundane things as the display technology, punters, and workstations. Though these technologies already play a major role in today's markets, I believe that their importance to scientific and engineering research and development to the evolution of information technology will be profound. The synergy between