Keeping the U.S. Computer Industry Competitive: Systems Integration (1992)

Now into its fifth decade, the so-called computer age has wrought an expectation of rapid technological progress. In fact, the sustained pace of rapid advance in hardware has become formulaic. The storage capacity of dynamic random access memory chips quadruples about every four years, processing speeds increase at double-digit annual rates, and the cost of computing declines between 20 and 30 percent each year. Add to these computing trends a growing array of digital communication devices and related technologies, and the raw capabilities of modern information technology can astound. Indeed, the technology—from automatic teller machines to laser scanning at the checkout stand to desktop publishing and to global distributed computing networks—has transformed spheres of business and many other areas of human endeavor. To many, today's applications are early manifestations of a seemingly endless stream of possibilities flowing from the font of information technology.

Still, society's appreciation, mastery, and application of information technology have grown more slowly and more fitfully than have the power and potential of hardware. The learning curve associated with the technology has proved to be steep. But organizations that have managed to move high up on the curve are reaping substantial benefits. These organizations have advanced beyond using computers to automate existing processes and have proceeded to enterprise-wide systems integration. Capitalizing on the convergence of computer and communications technology, they use hardware and very complex software to connect islands of computer applications into coordinated distributed network systems that often are global in reach. The integration entails far more than systems engineering—the physical linkage

of disparate and once-incompatible technologies. It involves melding application domains, such as manufacturing, finance, retail, or transportation, with a supporting infrastructure of information technology components, including databases, operating systems, architectures, networks, communications devices, and security measures. The result is a transformation of the business environment and of business processes, a strategic change aimed at rapid capture and assimilation of information for planning and decision making.

Today, executives and researchers speak of ''learning organizations'' and of the need for businesses to "reinvent" themselves as competitive conditions change. Both concepts reflect the critically important enabling role played by information technology.

Systems integration represents, in effect, the fruitful culmination of computing and communications research and experience, and it is digital technology's point of departure into the information age. "If there is any trend that has been consistent over the last four decades of computation, it is integration," explained Laszlo Belady, former vice president for software technology and advanced computing technology at the Microelectronics and Computer Technology Corp. (he is now chairman and director of Mitsubishi Electric Research Laboratories Inc.). "Forty years ago," he continued, "everything was isolated in terms of computer use—isolated in space and time. Now we see that both dimensions are integrated. There is continuity in electronic information, in time, and more and more applications have become interconnected. This leads to distributed computing networks."

Some U.S. businesses and government organizations are investing heavily in their information futures. Corporate spending for private networks now accounts for more than half of all investment in U.S. communications networks.¹ This has spawned an amalgam of uncoordinated networking initiatives, but many organizations remain untethered to the electronic connections that have sprouted over the last decade. Some small and midsize companies have formed information partnerships and pooled their resources to overcome the large financial and technical obstacles to exploiting the advantages of integrating and linking their information technology.² Most have not.

Many observers believe that the United States must build a national information infrastructure and that this information-age equivalent of the national highway system will not materialize from the ad hoc networks that have emerged in recent years. Advocates, including members of the Computer Science and Telecommunications Board, believe that the potential benefits wrought by new digital technology—as well as those to come, when information in all its diverse textual, graphic, and audio forms can be transmitted electronically—will multiply if the economy and all of society are integrated into a cohesive nationwide network with links to the rest of the world.³

In January 1991, the Computer Science and Telecommunications Board (CSTB) of the National Research Council sponsored a colloquium to contemplate the future of systems integration on scales small and large. Participants included representatives of manufacturers of information technology, systems integration firms, and businesses that are heavy users of information technology; government agencies; and universities.

In particular, participants evaluated the implications of the trend toward distributed computing networks for the U.S. computer and telecommunications industries. More generally, they assessed the significance of this trend for the U.S. economy and debated the roles of government, industry, and universities in developing a nationwide information infrastructure, a computing and communications network with universal access.

This report summarizes the proceedings of the day-long colloquium. Supplementary materials, referenced in the text, were used to augment or amplify the points made by participants. The remainder of this chapter provides a historical overview of the systems integration industry and of the trends underlying the growth of distributed computing networks. Chapter 2 describes applications of integrated information systems in business, as well as the challenges that information networking poses. Enabling technology, especially software and communications technology, is the focus of Chapter 3. In Chapter 4, views of the future and the evolution of today's information networks into systems of systems are discussed. The final chapter summarizes the main issues and themes that emerged during the colloquium.

Like the computer and the telephone, systems integration is a product of U.S. innovation. Its roots can be traced to the 1960s, when federal agencies first contracted with firms to design large-scale systems for data processing, communications, and aerospace and defense applications. Opportunities in the commercial sector grew in step with advances in computing technology and with the proliferation of proprietary hardware and software products. Some business customers persisted with a single vendor, assuming, sometimes incorrectly, compatibility across different families and generations of products. Requiring features or applications not offered by the supplier of their original equipment or responding to better price offerings, other customers picked and chose among rival vendors and became the owners of collections of incompatible hardware and software items. In addition, many organizations developed their own customized software applications. With each passing year, it became increasingly difficult to incorporate new technology into the existing base of hardware, software, and databases.

User demand for connectivity of hardware fueled the growth of systems engineering businesses, as did the emergence of affordable desktop ma-

chines, which led to what has been described as the democratization of computing. Subsequent demand for interoperability of software applications created new business opportunities. From these beginnings and with the aid of the computer industry's gradual and still-continuing migration to open architectures and networking standards, systems engineering evolved into systems integration, a more robust and more comprehensive undertaking that goes beyond making incompatible machines "talk" to each other. Today, systems integration is recognized as a problem-solving activity that entails harnessing and coordinating the power and capabilities of information technology in ways tailored to meet a customer's well-defined set of needs. The result is a complex one-of-a-kind system that, for a business, is intended to increase productivity, flexibility, responsiveness, and, ultimately, competitive advantage. Value is derived from the ability to achieve systemwide synergy among the technological elements and among the people linked by the network.

Typically, the end products are "large complex applications" built with a "broad range of knowledge," explained Ivan Selin, then under secretary for management in the U.S. Department of State (he is now chairman of the Nuclear Regulatory Commission). ''Although hardware and software may be the bulk of what is delivered, there is a lot of consulting, a lot of modification, interfacing, coding, installing—making this operation work in a large application," he said. Such work requires an interdisciplinary approach that includes not only hardware and software experts but also experts in the application domain for which the system is being developed (e.g., in a steel foundry it is vital to include individuals who understand steel manufacturing). According to Selin, founder of the systems integration firm American Management Systems, ''know-how" is the primary source of value added.

Unlike some other segments of the U.S. computer sector, the systems integration industry is thriving and is positioned well ahead of still-embryonic foreign competition. In 1990, the systems integration industry had estimated revenues of $14.9 billion; revenues in 1991 were projected to grow 13 percent to $16.8 billion.⁴ Enlarging foreign markets, especially Europe, are expected to drive rapid revenue increases during the next five years.⁵

About 1,600 firms earn all or the bulk of their revenues from systems integration. The largest of these include Electronic Data Systems, a subsidiary of General Motors with annual revenues of almost $6 billion, and Andersen Consulting, a division of the accounting firm Arthur Andersen & Co. Among the newest competitors in the commercial market are defense firms that have developed their systems integration expertise in designing

and developing complex computer-controlled weapons systems or parts of the U.S. military's sophisticated command, control, communication, and intelligence network.

In an industry traditionally structured to sell hardware components and supporting software products, systems integration is having profound effects on business strategy. Companies are maneuvering to meet the demands and needs of droves of customers who are linking their information technology into networks. In 1990 alone, U.S. businesses connected nearly 4 million personal computers to local networks, bringing the total number of networked PCs to more than 11 million.⁶

Many companies already providing computing and communication products and services have pegged systems integration and networked computing as key growth areas. Examples include the International Business Machines Corp., American Telephone & Telegraph Co., Digital Equipment Corp., Xerox Corp., Hewlett-Packard Co., Unisys, Motorola Inc., Compaq, and the NCR Corp. (which recently merged with AT&T). Major independent software publishers—among them Microsoft, the Lotus Development Corp., Computer Associates International, and the Oracle Corp.—have followed suit and have made networking applications a priority. Also vying in the market are retailer Sears, aircraft manufacturers McDonnell Douglas and Boeing, and accounting firms Arthur Andersen & Co., Price Waterhouse, Ernst and Young, and KPMG Peat Marwick. These and other firms with considerable experience in building and managing complex systems of information technology to support their main lines of business have created separate divisions to market their systems integration expertise.

Sometimes called flexible partnerships, alliances among firms are common. On one project, companies with complementary expertise may combine their resources and skills, but on another former partners may vie as competitors. Moreover, large businesses with in-house personnel assigned to building and managing information systems may call on outside firms for needed expertise in particular areas.

The strategic importance of systems integration and prospects for continuing high rates of systems integration are influencing both the behavior and structure of the computer and communications industries. In increasing numbers, hardware and software manufacturers are moving away from building proprietary equipment and applications and committing themselves to "open systems"—making software and hardware that conform with industry-accepted standards and, therefore, can work with the offerings of other vendors.

Recent acquisitions provide further evidence of the profound influence of systems integration and information networking. The most notable, perhaps, is AT&T's purchase of the NCR Corp., the nation's fifth-largest

computer manufacturer. A prime motivation behind AT&T's prolonged but eventually successful takeover, AT&T Chairman Robert E. Allen was quoted as saying, was to create a company capable of providing "global networks as easy to use and as accessible as the telephone network is today."⁷ Also acting to strengthen its competitive position in the market for networking applications and services, the Lotus Development Corp., one of the country's largest independent software publishers, purchased cc: Mail, a company with expertise in electronic mail applications. Previously, Lotus and Novell Inc., the largest maker of software for managing networks of personal computers, had explored the possibility of a merger, another indication of how companies are maneuvering to compete in the growing market for intra-and inter-networking products and services.

Systems integration remains largely an American phenomenon, but one that has captured the attention of foreign businesses—as customers and as competitors. Although beginning from a small base, foreign sales represent the fastest-growing market segment for some established U.S. systems integration businesses. Opportunities in foreign markets are likely to stimulate local competition, as well as draw new international competitors. Several colloquium participants said that they anticipated greater interest from Japanese electronics and telecommunication conglomerates, such as Toshiba, Fujitsu, NEC, Nippon Telephone and Telegraph, and Hitachi, each with annual revenues in excess of $10 billion. In Europe, cooperative government-sponsored research and development projects focusing on future computing and telecommunications technologies raise the specter of greater local competition in that lucrative market.

Despite the expectation that foreign firms will make a determined push in global markets for systems integration services, Belady and other colloquium participants suggested that at this stage in the evolution of information technology, the emerging opportunities are there for U.S. firms to lose. Systems integration, said Selin, is a "particularly strong example of [a] uniquely American ability to solve one-of-a-kind problems or solve very complex, very difficult problems in ways that are very hard to imagine. . . .[W]e're very effective when it comes to providing technological solutions to difficult problems."

For now, Selin added, the U.S. systems integration industry benefits from demonstrable successes that have made it the "recognized world leader." Because of U.S. accomplishments in designing, developing, and implementing large computer and communications systems, "being an American is a big advantage, not a disadvantage [as it is in] a number of other parts of our industry," he said.

Charles S. Feld, vice president for management information systems at Frito-Lay Inc., also noted a significant gap between the systems integration capabilities of U.S. companies and those of foreign firms. Having built an advanced information network that links all elements of its domestic operations, Frito-Lay intends to build a counterpart for its expanding international business, which has units in 22 countries. The task has proved to be very challenging. In his visits to foreign operations acquired by Frito-Lay, Feld has found that, "even though the hardware is exactly the same as what we would be using here, it is like being back in the 1960s in terms of the way people do projects and the way they think about things. I think we really do have a national asset here that is tremendous."

Nevertheless, there is concern that the flagging competitive status of U.S. firms in "upstream" hardware manufacturing industries, particularly semiconductor manufacturing equipment and certain segments of the semi-conductor industry, could place the U.S. systems integration industry at a disadvantage in international markets. Foreign firms with strengths in these building-block industries, such as the Japanese firms mentioned above, could leverage their technological advantages in the expanding market for computer and communications networks, especially if U.S. systems integration firms do not have timely access to new hardware products.

U.S. leadership in systems integration is attributed to a variety of factors, some more tangible than others. An acknowledged source of advantage is the continuing strong performance of the U.S. software manufacturers, who account for more than 60 percent of the global market. Software, particularly customized programs, is the primary source of the "value-added or content of the systems we build," explained Samuel H. Fuller, Digital Equipment Corporation's vice president for research. Virtually all colloquium participants agreed that maintaining the U.S. advantage in software design and development is critical to future success in the emerging global market for systems integration services.

"Ultimately," said Alfred V. Aho, assistant vice president at Bell Communications Research, "software is going to be the limiting factor" in determining the value and utility of integrated systems, as well as the success of firms vying to build these systems.

Also important but often unappreciated, according to several speakers, is the advantage conferred by international acceptance of many computer and communications standards that originated in the United States. This conformance with U.S. standards reflects in large part the nation's pioneering role in developing and applying information technology, resulting in a kind of after-the-fact acceptance of proven technologies. International adoption

of technology developed in the United States opened world markets to U.S. firms and obliged local competitors to follow the U.S. lead.

However, international standards organizations are becoming much more proactive and are attempting to anticipate technological progress. Their aim is to integrate important innovations into standards as new technology nears commercial application, avoiding the ferment of de facto standardization and providing information-technology users with an efficient means of linking existing and new hardware and software. In practice, international standards bodies have far to go to achieve this model of efficiency, but proactive approaches significantly enhance the strategic importance of standards making. Firms that are the most adept in the standards arena may be able to leapfrog the current competition when new generations of computing and communication technology are ready for the marketplace. In the opinion of some colloquium participants, U.S. firms and the federal government have been far more passive than their European and Japanese counterparts in international standards-setting activities that are likely to shape the future of information networks.

The success of U.S. systems integration firms may also stem from what a number of colloquium speakers described as an American capacity for flexible thinking and managing complexity, a hallmark of integrated information systems. Irving Wladawsky-Berger, assistant general manager of development and quality for IBM's Enterprise Systems, emphasized that the nation's prowess in design and innovation and its propensity for change helped stake the U.S. industry to its leadership position. Selin expressed a similar view. "As long as our computer industry concentrates on building applications and building problem-solving systems and not just trying to produce chips or simple devices in larger numbers and more cheaply than other people do, I think we'll do very well," he said.

But the complexity wrought by an ever-growing web of interconnected technologies, people, and organizations has also led to efforts to create order and simplicity. Standards are one manifestation of these efforts (see Chapter 3 appendix, "Standards Making at a Glance"). For example, the evolving standard for Integrated Services Digital Network (ISDN) service promises to provide an interface for worldwide networking. Growing interest in object-oriented programming and in reusable software elements is another illustration of efforts to simplify systems integration and eliminate some of the complexity and redundancy in building information networks.

Within these emerging areas, cautioned Wladawsky-Berger, may lie the seeds of problems for U.S. systems integration firms. "Executing well is absolutely paramount," he said. "It may be that the design part of systems integration, which is where we excel, is becoming an ever-smaller part of the total problem, and the execution part of it—that is, 'get the damn thing built and working'—is where more and more of the problems are. Now,

notice that, while we in the United States are superb in this ferment of design and innovation, lots has been written about our weaknesses in the back end, in executing. . . . That is something I think we need to watch out for."

Selin noted, however, that greater standardization and uniformity in applications would likely create more opportunities, at both the low and high ends of the market. With increasing standardization, he said, "the detailed specification of the components becomes less important. . . . What happens is that for problems that took a one-of-a-kind solution, where everything was customized and therefore so expensive that very few companies could afford to solve their problems, it now becomes feasible to do three-quarters of the software with an applications-specific piece of software. Therefore," he continued, "you make it economical for a much larger range of companies to come in. . . . [T]he total size of the market will increase considerably because more and more organizations will be able to afford what only a few could before. At the same time, you go out and you do tougher things."

As technology and skills improve, Selin added, systems integrators will be called on to solve "one-of-a-kind problems that are more complex than anything we thought possible 10 years ago."

A common base of digital technology is evolving rapidly. The computer now exists in many forms—from supercomputers and massively parallel machines to the ubiquitous desk-top and lap-top personal computers to the highly portable pen-based and palm-top computers that are now emerging. Just as significant as this diversity is the tremendous variety of complementing devices for communicating, gathering, and presenting information. Besides the familiar printer, the "peripheral" devices of computers now include copiers, facsimile machines, telephones, answering machines, pagers, scanners, electronic cameras, data-storage devices, optical and electronic sensors, and a vast and growing array of computer-controlled machinery.

Although the decade-old prediction of a personal computer in every home is not even close to being realized, digitally controlled devices with embedded software and hardware are pervasive in the home: video cassette recorders and cameras, compact disc players, telephones, microwaves, and other appliances. In prospect is an all-digital high-definition television system, a development that could transform that most ubiquitous and passive of household appliances into a machine capable of storing, retrieving, and manipulating video images.

The convergence of technologies is also evident in the area of wireless communications, now principally the domain of cellular telephones. Several

firms have developed wireless methods for linking computers in networks. Such over-the-air connections between computers will make network configurations as flexible as those for cellular telephones.

As the communication capabilities of computers grow and as communication devices become more computerlike, opportunities for information services and products multiply. The hybridization of information technology blurs distinctions between industries, giving rise to rivalries among firms that once competed in separate markets. For consumers, the ubiquity of digital technology creates exciting new opportunities—new ways to gather and manipulate information, new ways to work and recreate, and new ways to collaborate. For businesses, the technology creates new opportunities that extend throughout and across organizations. Many firms are now seeking to achieve the synergies afforded by information networking through systems integration.

Today, however, systems integration is an option largely restricted to large firms, some of which have invested over $1 billion to revamp their corporate information systems and create an infrastructure for distributed computing networks. Such high costs attest to the complexity and the magnitude of the task. The expense and complexity also reflect the difficulties posed by the lack of standards and communication protocols essential for networks linking a heterogeneous collection of equipment. As a result, many of today's networks do not achieve the ideal of a seamless computing and communications environment. Each brand of equipment may have its own control computer, and even the most vigilant network management teams can have trouble isolating and responding to problems that can disrupt the entire information system. The potential for problems is magnified, and new ones, such as the threat of security breaches, arise when organizations link their private networks to those of other organizations or when network access can be achieved through public telephone lines.

As ambitious as they are, today's systems integration projects are microcosms of the nation-and globe-spanning information networks that are expected to evolve. In effect, said Robert L. Martin, vice president for software technology and systems at Bell Communications Research, the benefits now being reaped by firms that have invested in their own information networks would be extended to small businesses and individuals. "You could bring intelligence further down the economic hierarchy," he added.

Visions of the computer age's more revolutionary successor, the information age, cast the computer as a gateway that provides access to information in virtually all parts of the world. Proponents of the information-age

equivalent to universal telephone access contend that a national information infrastructure is vital to the nation's economic competitiveness.

"I believe the broader issue is to change the concept of a system to a highly networked distributed base," Martin said. "You would then have an education network that could bring the appropriate educational talent to the disabled, the gifted, or the average student. You would have a very different form of health system."

Such thinking is not confined to the United States. In Europe and Japan, government and industry are funding collaborative research and development projects to address the technological hurdles that must be overcome to develop an infrastructure and the services it would support. The Japanese and European governments are also fostering development of key standards for achieving compatibility and interoperability. Several European nations have implemented accelerated schedules for installing the digital networks and common channel signaling necessary for ISDN service. In Japan, the Nippon Telephone and Telegraph Co. has announced its intention to develop the most advanced telecommunications system in the world. By 1995, it intends to offer ISDN service in all 56 Japanese cities, and by the year 2015, it will link all homes by optical fiber.⁸ At the same time, Japan's Ministry of International Trade and Industry (MITI) has initiated a series of programs designed to integrate its computer, telecommunications, and other information-related industries and to network its society. Recently announced initiatives include research and development projects on interoperable database systems and improved graphical user interfaces, modeling studies of "information age cities," and several cooperative research programs aimed at helping developing countries build their own information networks.⁹

Colloquium participants did not see U.S. government and industry as attending to the nation's information infrastructure with the same levels of determination and comprehensiveness as those exhibited in Japan and Europe. Virtually all endorsed administration and legislative proposals calling for the creation of a National Research and Education Network (NREN), an information "superhighway" that eventually would transmit data at rates of several billion bits per second, or more than 1,000 times faster than today's standard data networks. As proposed, NREN would connect the nation's universities and collaborating companies. The administration's initiative, called the High Performance Computing and Communications (HPCC) program, would provide a focal point for developing the hardware, software, and systems that a truly national information infrastructure will require. (See Chapter 4 for a discussion of the HPCC program.)

Indeed, whether practiced on the scale of a single company or on the scale of an entire nation, systems integration poses a variety of technical, social, economic, and regulatory issues, which are examined in the next chapters.

1. Durros, Irwin. 1990. "Calling for Cooperation," Bellcore Exchange, November–December, p. 7.

2. For example, 18 midsize paper companies with collective annual revenues of $4 billion pooled their resources to create a $50 million global information network that links the companies, offices of major customers, and international sales offices. This system permits same-day responses to inquiries from customers, as compared with the industry average of 12 days. (Konsynski, Benn R., and F. Warren McFarlan. 1990. "Information Partnerships—Shared Data, Shared Scale," Harvard Business Review, September–October, pp. 115–118.)

Also, a number of Wall Street security firms—where information technology expenditures can account for 15 percent to 20 percent of expenses—are engaged in several types of cooperative activities, including sharing disaster-recovery sites, combining and distributing their analytical libraries, and developing a shared Electronic Data Interchange network. (Ambrosio, Johanna. 1991. "Wall Street Firms Try Shared Technology," ComputerWorld, June 24.)

3. Computer Science and Technology Board, National Research Council. 1988. The National Challenge in Computer Science and Technology, National Academy Press, Washington, D.C. See also the September 1991 issue of Scientific American, which has several articles relating to information infrastructure.

4. Department of Commerce. 1991. U.S. Industrial Outlook 1991, Information Services, Government Printing Office, Washington D.C., p. 27-5.

5. As reported in Gartner Group Inc. 1990. "Systems Integration Scenario," Gartner Group Inc., Greenwich, Conn., p. 7.

6. Rothfeder, Jeffrey, Peter Coy, and Gary McWilliams. 1990. "Taming the Wild Network," Business Week, October 8, p. 143.

7. Shapiro, Eben. 1991. "AT&T Buying Computer Maker in Stock Deal Worth $7.4 Billion," New York Times, May 7, p. A1.

8. Bellcore Technical Liaison Office. 1990. "Japanese Telecommunications," unpublished paper, Bellcore, Morristown, N.J., July, p. 2-1.

9. For a compelling discussion of the enormous potential for an information infrastructure to transform society as seen by Japan's Ministry of International Trade and Industry (MITI), the reader is referred to the so-called "Sixth Generation" or New Information Processing Technology (NIPT) project currently proposed by MITI and discussed in its Report of the Research Committee on New Information Processing Technology, Industrial Electronics Division, Machinery and Information Industries Bureau, Ministry of International Trade and Industry, March 1991.