Design in the New Millennium: Advanced Engineering Environments: Phase 2 (2000)

America is changing. Many of the most noticeable changes in day-to-day life are associated with the advancing capabilities of computer systems, the growing variety of tasks they can accomplish, and the accelerating rate of change. This report addresses the development of advanced engineering environments (AEEs) over the next 15 years.¹ Inherent in this charge is the assumption that the future of computer applications and technology is foreseeable on a 15-year time scale. Experience has shown, however, that reality usually defies predictions. In some cases, predictions turn out to be overly optimistic, as with past expectations about the future of artificial intelligence or intelligent highways. In other cases, predictions have been greatly exceeded. For example, in 1989 supercomputers cost between $1 million and $20 million, and the best commercially available machines, such as the Cray Y-MP, could perform about 1 billion (giga) floating-point operations per second (i.e., 1 gflop) using 8 interconnected central processing units.

In 1989 a National Research Council report predicted that in the following 10 years general-purpose supercomputers would demonstrate 100 billion gflops using from 32 to 256 interconnected central processing units (NRC, 1989). By November 1999, the most capable supercomputer in the world could perform 2,400 gflops using 9,632 linked processors, and commercially available machines, such as the Cray SV1, could be configured to provide up to 1,000 gflops. The National Research Council report also predicted that the cooling requirements for a supercomputer the size of a suitcase would be so great that there would be an instant meltdown if the cooling system failed. Today, the specialpurpose computer processing unit in the PlayStation^® 2 video game console is capable of 6.2 gflops, yet it is the size of a typical video game and has no cooling system because it consumes just 15 watts of power.

The National Research Council and National Academy of Engineering identified many barriers to the widespread use of AEEs in the Phase 1 report of the Committee on Advanced Engineering Environments (NRC, 1999). Key barriers included incompatibility of software and hardware; information management systems incapable of dealing effectively with vast amounts of data; the lack of definable metrics to justify expenditures on AEEs; and the need to integrate AEE education and training into academic programs. In this Phase 2 report, the committee describes longterm approaches that industry, government, and academia could use to achieve the AEE vision.

Large companies (such as Boeing and DaimlerChrysler) and small companies (such as Concepts, ETI, a designer and manufacturer of turbomachinery) have demonstrated the costs and benefits of pioneering the use of AEE technologies and systems. Initial costs can be high, but AEEs can initiate a rising tide of improvements in terms of shorter development times, reduced development costs, and improved product performance. Continued improvements in the engineering enterprise will depend partly on continued growth in the capabilities of computational and communications systems. Improvements in interactive communication technologies will create many new opportunities for engineering collaborations via distributed networking, telecommunications, multiuser computer software, and interactive virtual reality. Interdisciplinary collaborations will be especially important for implementing comprehensive processes that can integrate the design of mechanical systems with the design of electrical systems and software. Successful collaborations, however, will require first overcoming incompatibilities between emerging technologies and the existing technological infrastructure and organizational cultures.

The Phase 1 report provided initial guidance on the general approaches and roles for industry, government, and academia in pursuing more advanced and capable AEEs. The additional information collected for the Phase 2 report confirmed that the recommended approach would provide a solid foundation for achieving long-term goals. Key steps in this top-level process include the following:

• forming a national partnership of government, industry, and academia to take advantage of the current historic opportunity to develop AEEs

• forming government-industry-academia AEE partnerships by individual agencies, such as the National Aeronautics and Space Administration (NASA), as an interim step for addressing agency-specific goals while a national partnership is being formed

• overcoming major barriers related to the integration of systems, tools, and data; information management; cultural, economic, and management issues; and education and training

• facilitating the transfer of new capabilities to commercially available products by developing applicationspecific tools required by government through contracts with industry whenever practical

• focusing the govemment's AEE research and development on key objectives, such as (1) modeling key physical processes, (2) improving generic AEE methodologies and automated tools, (3) developing testbeds that simulate user environments, (4) developing accurate performance metrics, and (5) other areas where market-based incentives are not motivating adequate industry-sponsored research

• providing government incentives for (1) industry to adopt AEE technologies in government procurements, (2) academia to adopt AEE technologies in major government-sponsored research programs, and (3) industry and academia to collaborate in modernizing educational curricula to prepare students for an AEE work environment

Government agencies involved in research and development of AEE technologies should be more aggressive in forming a national partnership with industry and academia to develop AEEs that offer seamless, end-to-end engineering design capabilities encompassing the entire life cycle of products and missions (see Table ES-1). The federal government is spending more than a billion dollars each year on research for advanced computing technologies, and industry is investing even more. As a result, most of the computational and communications technologies needed to create AEEs will probably be developed, even if the needs of AEEs are not directly considered. Building new AEEs that take advantage of these new technologies, however, will be impeded if AEE requirements are not considered during technology development. For example, small adjustments in some future Internet-related technologies and applications (which are discussed below) might make a big difference in their ability to support interoperable AEEs. Areas of particular interest include latency, quality of service (i.e., the ability to guarantee users that they will have the communications bandwidth necessary to conduct AEE operations at a given time), and security (i.e., the ability to provide authorized users with easy access to data and systems, while protecting competition-sensitive data from unauthorized disclosure or alteration). For example, widespread use of AEEs would involve simultaneously providing many users with reliable, low-latency, high-bandwidth service. A national partnership that includes developers of AEEs and Internet-related technologies and applications should be formed to develop (1) open architectures and functional specifications for AEEs, (2) plans for transferring the results of government research and development to the commercial software industry and software users, and (3) approaches for resolving information management and organizational issues.

An AEE is not just a technology system. The purpose of AEEs is to improve the efficiency with which large, distributed teams design and implement large, complex systems. To achieve this goal, the national partnership will have to address psychosocial issues related to human-machine

interfaces and the manner in which work is carried out by the people who use AEE technologies and systems.

NASA is developing AEE-related technologies and has facilities in several different regions of the United States. NASA also has a wide-ranging mission, which includes encouraging “the most effective utilization of the scientific and engineering resources of the United States, with close cooperation among all interested agencies” (National Aeronautics and Space Act of 1958, Public Law 85-568, as amended). Furthermore, many NASA programs challenge the limits of human ingenuity and involve leading engineering universities and corporations. Thus, NASA would benefit from increased use of AEEs, and it is well positioned to form a government-industry-academia partnership for NASA missions as an interim step (while a national partnership is being formed). However, the conditions necessary for significant, widespread adoption of AEEs in the agency do not yet exist. Sustained leadership, high-level organizational commitment, adequate funding, and a cohesive plan that includes all NASA centers is necessary to accelerate the adoption of AEEs within NASA.

NASA should not create a broad-based AEE research program to develop comprehensive AEE systems. Instead, NASA should advocate and facilitate greater use of AEE technologies by contractors involved in NASA programs, capitalize more on commercial technologies, fund research and development to satisfy NASA's specialized needs, quantify and advertise the benefits of AEEs in different applications, and support advances in the state of the art in focused areas that industry's market-driven research and development are not addressing. NASA should also ensure that its systems are compatible with the AEE technologies and systems used by its research and development partners in industry, academia, and other government agencies. In particular, NASA should investigate how NASA-funded research related to AEEs and the Next Generation Internet can enhance the ability of Internet-related technologies and applications to meet AEE objectives.

Creating and maintaining an AEE-qualified workforce will require changes in education and training. AEEs could enable students to learn in virtual environments, foster the development of critical thinking skills, and, perhaps, improve educational efficiency so much that AEE-related education could be integrated into curricula without requiring the elimination of existing courses. AEEs could also be used to integrate schools of engineering with schools of liberal arts and sciences. Because the capabilities of AEEs are not limited to traditional engineering tasks, access to AEEs should not be limited to engineering students.

Just as individual corporations will require a champion to implement AEEs, engineering schools will need influential champions to integrate AEEs into the university environment and sustain support for AEEs and related interdisciplinary programs. Champions will be especially important if individual academic departments do not embrace AEEs. At many universities, strong internal leadership, combined with external pressure from accrediting organizations and industrial engineering organizations, will be necessary to encourage faculty to use AEEs and modify undergraduate and graduate curricula accordingly. In a broader context, the federal government can facilitate change by funding long-term interdisciplinary research associated with AEEs and, with industry, by including academia in a national partnership for fostering the development of AEE technologies and systems.

Internet-related technologies and applications include software and hardware that provide the basic capabilities of the Internet to (1) transfer large amounts of data quickly and reliably among interconnected public and private networks and (2) enable shared, distributed applications and objects (such as e-mail, Web browsers, Internet-based video-conference systems, and simulations) that enhance the functionality of the Internet to the general public, including the science and engineering communities. Internet-related technologies and applications may be associated with (1) the current Internet, (2) the Internet of the future (in whatever form it takes), or (3) current or future private networks, which may be established to provide a guaranteed level of quality of service or to test advanced technologies. The committee recommends that AEEs be designed in a way that is compatible with and takes advantage of Internet-related technologies and applications. However, AEEs are in a distinct class because they are intended for a special purpose and have a small market compared to the Internet as a whole. Thus, the development of Internet-related technologies and applications will proceed without reference to the needs of AEEs unless special efforts are made to integrate them.

Advanced Internet technologies are one of the keys to developing AEEs that can overcome critical technical, cultural, management, and educational barriers. One of the most difficult long-term barriers concerns integration of software tools for design and development across (1) disparate operating systems, distribution networks, and programming languages and (2) different governmental and corporate cultures. Overcoming this barrier in the next 15 years will require general solutions to the problems of interoperability (i.e., the ability of various systems to work together in a meaningful and coherent fashion) and composability (i.e., the ability to build systems using components designed for other systems). The committee believes that the current move

toward the Internet as a universal medium will make it easier to solve these complex problems. Accordingly, the government's AEE research and development programs should greatly increase their emphasis on technologies, open standards, industry-wide consortia, and other processes that have contributed to the success of the Internet. AEEs should benefit as much as possible from Internet-related technologies and applications being developed by other research programs, such as the Internet-2, the Next Generation Internet, the very high performance Backbone Network Service, and the Information Technology Research Initiative.

New telecommunications and collaborative capabilities of future Internet-related technologies and applications will enable users in the same or separate locations to engage in interdependent, cooperative activities using a common computer-based environment. The nature and extent of these collaborations will depend, in part, on the extent to which AEE developers have been able to integrate their technologies with the Internet. Just as important, however, will be bringing mission-oriented communities together with the computer and social science communities to address cultural barriers. Interdisciplinary studies are necessary to develop AEEs that enable workers with various work styles to form effective teams and to accommodate the psychological and temporal dimensions of synchronous, distributed, collaborative activities, especially if they involve personnel in multiple time zones and organizations with different cultures and business goals. Coordinating the development of AEEs with the development of Internet-related technologies and applications, such as tele-immersion capabilities, would increase the likelihood that solutions to cultural barriers will be incorporated into the technologies that underlie the future Internet. This coordination could also lead to AEEs with a feel similar to the future Internet, thereby reducing the need for specialized equipment and training.

NRC (National Research Council) . 1989 . Supercomputers: Directions inTechnology and Applications . Computer Science and Technology Board. Washington, D.C. : National Academy Press . Available on line at: http://www.nap.edu/catalog/1405.html May 3, 2000 .

NRC . 1999 . Advanced Engineering Environments: Achieving the Vision (Phase 1) . Committee on Advanced Engineering Environments . Washington, D.C. : National Academy Press . Available on line at: http://books.nap.edu/html/adv_eng_env/ April 5, 2000 .