National Academies Press: OpenBook

Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment (2009)

Chapter: 3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007

« Previous: 2 Information Technology: The Essential Enabler for the Information Society
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 42
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 43
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 44
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 45
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 46
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 47
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 48
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 49
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 50
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 51
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 52
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 53
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 54
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 55
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 56
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 57
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 58
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 59
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 60
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 61
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 62
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 63
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 64
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 65
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 66
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 67
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 68
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 69
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 70
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 71
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 72
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 73
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 74
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 75
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 76
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 77
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 78
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 79
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 80
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 81
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 82
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 83
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 84
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 85
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 86
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 87
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 88
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 89
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 90
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 91
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 92
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 93
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 94
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 95
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 96
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 97
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 98
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 99
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 100
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 101
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 102
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 103
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 104
Suggested Citation:"3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007." National Research Council. 2009. Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment. Washington, DC: The National Academies Press. doi: 10.17226/12174.
×
Page 105

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

3 The Changing Landscape of the U.S. Information Technology R&D Ecosystem: 1995-2007 This chapter reviews the evolution of the information technology (IT) research and development (R&D) ecosystem in the time period 1995 through 2007. As with any ecosystem in nature, the IT R&D ecosystem responds to external forces to which it has been subjected, in turn influ- encing those forces by the way that the system evolves. The time from the mid-1990s to the present has been a period of almost unprecedented change, in the global, technical, and industrial contexts. The chapter is organized in four sections. The first reviews the shocks to the U.S. IT R&D ecosystem in terms of the rise and aftermath of the speculative financial bubble. The second section discusses the emergence of new technology platforms, based on open-source software, collab- orative community development, and Web-centric technologies, and the challenges that these present to traditional IT industrial organization. The third section addresses the rapid globalization of the underlying IT industrial sectors, with a particular focus on the cases of the semiconduc- tor, computer, and software industries. It also describes the rise of new regions where IT R&D is performed, both nationally and internationally, focusing particularly on the new IT powerhouses of India, China, and Taiwan. The fourth section describes the role that infrastructure plays in enabling innovation and the importance of enhancing U.S. broadband local-access infrastructure. The chapter concludes with a brief summary. Shocks to the U.S. Ecosystem The period 1995 to 2007 has been a turbulent one for the U.S. economy and the world economy. The early part of the period was characterized 42

The changing landscape: 1995-2007 43 by an IT-fueled speculative boom, followed by an economic bust from which the IT sector is just now starting to emerge. This section reviews the shocks to which the IT R&D ecosystem was subjected during this time. The Rise of “Irrational Exuberance” The term “irrational exuberance” is credited to former Federal Reserve Board Chairman Alan Greenspan. In a speech given at the American Enterprise Institute in 1996, Greenspan made a general observation about the difficulty of recognizing “unduly escalated asset values.” Few seemed concerned about the possibility of irrational exuberance or unduly high asset values as the excitement about the Internet created the dot-com boom. The initial public offering (IPO) of Netscape Communications Cor- poration in August 1995 symbolizes the beginning of this period. With the benefit of hindsight, those early years can be seen to have fueled a mas- sive expansion and upgrade of the global telecommunications system and powered the adoption of the Internet, but when the bubble of excitement burst in 2000, a powerful jolt was inflicted on the IT R&D ecosystem. The introduction and adoption of the World Wide Web were predi- cated on the ubiquity of the personal computer. With intuitive browsers and simple mark-up language, the Web enabled millions of individuals and businesses to create Web sites that could reach hundreds of millions of people and to engage in commerce. Companies formed rapidly, raising venture capital at high valuations to pursue new Web-enabled opportuni- ties. Entrepreneurs and investors were lured into adopting business plans with weak fundamentals and (at least in retrospect) objectives that did not create lasting and tangible customer value. And yet, despite the agony of the bursting bubble, the Internet changed the lives of hundreds of millions of people around the world. Figure 3.1 shows the rapid growth in venture funding for IT start-ups, particularly in the software and telecommunications sectors, following the creation of the Web in the mid-1990s. (For comparison, biotechnology funding, which did not experience such extreme funding changes, is also shown.) According to data from the MoneyTree survey, the number of IT start-up companies receiving venture investments reached a peak of more For a broad study of the 1995 stock market boom in the context of others and with respect to structural factors contributing to speculative bubbles, see Robert Schiller, Irrational Exuber- ance, Princeton University Press, Princeton, N.J., 2000. The overall peak in terms of both the total number of venture deals and total amounts raised came in the first quarter of 2000: in that quarter there were 2,129 deals amount- ing to $28,414 million, according to PricewaterhouseCoopers/Thomson Financial/National Venture Capital Association MoneyTree Report historical data, available at https://www. pwcmoneytree.com/MTPublic/ns/nav.jsp?page=historical; accessed August 20, 2007.

44 assessing the impacts of changes in the it R&D ecosystem Software 30 Telecommunications Computers and 25 Peripherals Networks and 20 Equipment U.S. $ (billion) Semiconductors 15 Biotechnology 10 5 0 1995 1997 1999 2001 2003 2005 2007 Year Figure 3.1  Total amount of investments by year, 1995 through 2007, in venture Figure 3-1.eps funding for IT start-ups in five sectors, compared with that for biotechnology start- ups. The investment bubble in IT start-ups ballooned after 1995 and was deflating by 2001. Source: Data (national aggregate data by industry from 1995 to 2007) from PricewaterhouseCoopers/Thomson Financial/National Venture Capital Association MoneyTree Report. Available at https://www.pwcmoneytree.com/MTPublic/ns/nav. jsp?page=notice&iden=B. than 5,600 in the year 2000, compared with fewer than 1,300 in 1995; the average investment per deal was over $18 million for year 2000 invest- ments, compared with $4.5 million in 1995. The period 1995 through 2000 was characterized by a number of high- profile success stories along with many large-scale capital deployment mistakes. In retrospect, too many of the IT start-ups that received ven- ture capital funding were ill-conceived, too few of the funded start-ups had solid business fundamentals, and ultimately most squandered their invested capital and went bankrupt. Data from PricewaterhouseCoopers/Thomson Financial/National Venture Capital Asso- ciation MoneyTree Report, available at https://www.pwcmoneytree.com/MTPublic/ns/nav. jsp?page=historical; accessed November 20, 2008.

The changing landscape: 1995-2007 45 The financial woes, and even bankruptcies, of several large telecom- munications providers suggest that too many of them attempted to build next-generation data network backbones. This investment occurred even as the somewhat lackluster state of U.S. broadband suggests that too little investment went into the transformation of U.S. narrowband local-access infrastructure into a broadband local-access infrastructure—thus creat- ing a potential impediment to the deployment of cutting-edge applica- tions and services that depend on high data rates. (See the section below entitled “Infrastructure to Enable Multifaceted Innovation.”) At the same time, there are differing perspectives on the net success of investment strategies during this period. Researchers have begun to explore some of the entrepreneurial and investment dynamics of the era. For example, drawing on data contributed to the Business Plan Archive,  Goldfarb, Kirsch, and Pfarrer estimate that many less-visible and privately held ventures—accounting for nearly half of dot-com era ventures—sur- vived until 2004. Thus, they characterize the dot-com era as a “legitimate response to a technology shock.” However, these authors also recognize that “many good opportunities were oversold to investors and the public as large opportunities” and that the bursting bubble brought reduced, but more realistic, private and public market valuations. Subsequent research by Goldfarb, Kirsch, and Miller explored the implications of the pre-2000 “pervasive and persistent belief” in a “Get Big Fast (GBF)” business strategy that was based on the market preemp- tion of competitors and on expected economies of scale associated with network effects. They concluded that belief in the GBF strategy by entre- preneurs, investors, and the public led to overly focused investment in too The Business Plan Archive (www.businessplanarchive.org) was established in 2002 to preserve business plans and other digital ephemera from the dot-com era technology companies. Using a sample of new technology ventures drawn from the funding solicitations re- ceived by one venture capital fund, Goldfarb, Kirsch, and Pfarrer extrapolated estimates of venture creation during the 1996-2002 period that include transactions not published in the Thomson Financial data. They estimate that 50,000 new ventures were formed to exploit the commercialization of the Internet and that 24,000 of these received some $256 billion from formal and informal investors over the 1996-2002 period. By contrast, according to the authors, Thomson Financial reported only 8,500 transactions but the vast majority ($217 billion) of the investment during this period. See Brent D. Goldfarb, David A. Kirsch, and Michael D. Pfarrer, “Searching for Ghosts: Business Survival, Unmeasured Entrepreneur- ial Activity and Private Equity Investment in the Dot-Com Era,” Robert H. Smith School of Business Working Paper No. RHS-06-027, October 2005, available at http://ssrn.com/ abstract=825687; accessed December 1, 2007. Brent D. Goldfarb, David A. Kirsch, and Michael D. Pfarrer, “Searching for Ghosts: Busi- ness Survival, Unmeasured Entrepreneurial Activity and Private Equity Investment in the Dot-Com Era,” Robert H. Smith School of Business Working Paper No. RHS-06-027, October 2005, available at http://ssrn.com/abstract=825687; accessed December 1, 2007.

46 assessing the impacts of changes in the it R&D ecosystem few start-ups and resulted both in too little entry and in both private and public market overcorrections once its limitations were realized.  In retrospect, the GBF strategy had mixed results. It did not work well for a number of firms such as Webvan, Pets.com, or Scient, whose failures and resulting financial losses contributed to a negative public perception of the era. However, the strategy worked well in the long term (i.e., until the present) for a small number of firms such as Amazon, Yahoo!, eBay, and Monster. Indeed, the existence of an environment that led to the suc- cessful creation of these leading firms is arguably a unique strength of the U.S. IT R&D ecosystem. From another perspective, however, the rapid birth and death of IT start-ups and the venture investments during the peak amounted to a valuable form of experimentation—with business models, customer pref- erences, consumer adoption of the relatively new Internet for entertain- ment and commerce—that was both faster and perhaps more effective than applied research in a university environment on the same ques- tions might have been. For society as a whole, the hundreds of millions of “written-off” venture dollars went toward market experimentation and toward operationalizing the scientific and engineering advances made possible in part by traditional IT R&D. Nonetheless, going for- ward, one lesson from this period is that capital misallocation might be avoided through greater focus on how consumers use and value IT (see the section entitled “Infrastructure to Enable Multifaceted Innovation,” below). “Y2K” and the Development of the Indian Software Industry The year 2000 problem, known as Y2K, or the Millennium Bug, refers to the perceived difficulty of handling 21st century dates in older but still used computing systems. Industry and government voiced concern that mission-critical software that used only two digits to store years would confuse 2000 with 1900 when the calendar wrapped from 1999 to 2000. In the end, after considerable effort to revamp software or introduce opera- tional workarounds, the disruptions caused were minor. Global efforts to address the Y2K problem provided a major growth impetus for the Indian software industry. Few programmers in the United BrentD. Goldfarb, David A. Kirsch, and David A. Miller, “Was There Too Little Entry During the Dot Com Era?” Journal of Financial Economics 86(1):100-144, 2007, available at www.sciencedirect.com; accessed December 4, 2007. David Kirsch and Brent Goldfarb, “Small Ideas, Big Ideas, Bad Ideas, Good Ideas: Get Big Fast and Dot Com Venture Creation,” Robert H. Smith School of Business Working Paper No. RHS-06-049, November 2006, available at http://ssrn.com/abstract=946446; accessed December 1, 2007.

The changing landscape: 1995-2007 47 States were prepared, or even wanted, to deal with fixing the legacy software problems. Firms turned to Indian companies to develop tools to analyze their software for vulnerabilities and to modify their applica- tions. The build-out of global telecommunications services and adop- tion of the Internet that had recently taken place in India and the rest of Asia helped make it possible to coordinate activities between the United States and overseas. India’s Y2K success enticed global IT firms to locate development laboratories and product groups there, where products for global markets are being developed. By 2007, Indian firms had emerged as world-class IT services enterprises. NASDAQ Bust (2000) The NASDAQ—National Association of Securities Dealers Auto- mated Quotations—bust came in the aftermath of a soaring stock market of the 1990s. It peaked in March 2000, and the bust began when the valu- ations of firms with questionable business models could no longer be sustained. The immediate effect was to end the unsustainable expansion plans of many firms, leading to retrenchments and even bankruptcies. A recession ensued, even affecting IT workers. Venture capital investments declined, with start-up companies encouraged to outsource development to reduce costs. The public markets no longer supported technology IPOs, reducing the returns to early-stage investors and increasing inves- tors’ aversions. The general sense of pessimism about the IT sector and a perceived lack of employment appear to have led to the ensuing decline in enroll- ments in computer science programs. This happened at the same time that the software industry in India and Eastern Europe enjoyed high rates of growth, helping to accelerate the migration of projects to these regions. In 2007, the markets and the field continued their recovery. Venture capital investment was up, reaching the highest levels since 2001. There had been a number of spectacular recent IPOs, including those of Google (2004), Riverbed Technology (2006), and VMware (2007). The number of technology IPOs on U.S. exchanges was once again increasing,10 albeit at a sobered pace.11 As the pace of investment in IT firms increases, access to talent becomes a limiting factor. The demand for students with IT skills See Kristina Shevory, “In Silicon Valley, Steady But Cautious Growth Returns,” New York Times, June 27, 2007. 10See “Door Is Open to High-Tech Offerings That Meet Thresholds,” New York Times, June 29, 2007. 11As this report went to press in 2008, there were indications of at least a temporary falloff in IPOs, reflecting prevailing economic conditions.

48 assessing the impacts of changes in the it R&D ecosystem is increasing, but with their numbers reduced, salaries rise and firms look globally for technical talent.12 Aftereffects of September 11, 2001 The terrorist attacks of September 11, 2001, profoundly affected the United States, with a redirection of national attention focused on address- ing the terrorist threat. Subsequent conflicts in Afghanistan and Iraq led national resources to be redirected to the war effort. There is evidence that some IT research funding has been redirected to national and home- land defense objectives.13 This type of research is not easily performed in universities owing to campus restrictions on classified research and the presence of large numbers of students, faculty, and other researchers who are foreign nationals. Funding data suggest that the Department of Defense (DOD) has reduced its investment in university research programs.14 Furthermore, the post-September 11 environment had other effects. Anecdotal evidence suggests that in the immediate aftermath of Septem- ber 11, foreign students found it more difficult to enroll in U.S. graduate programs owing to visa difficulties and new background checks at U.S. embassies and consulates abroad.15 In any case, survey data were col- lected in early 2004 by the Council of Graduate Schools (CGS) from the 113 graduate schools that enroll nearly half of all international graduate students in the United States. These data indicated an overall decline of 12The average salary offer for a college graduate with a computer science major was $53,051 in 2007, up 4.5 percent from the previous year. Only graduates with majors in chemi- cal, electrical, and mechanical engineering had higher average starting salaries. See National Association of Colleges and Employers data reported by Computing Research Association, available at http://www.cra.org/wp/index.php?p=123; accessed February 20, 2008. 13See John Markoff, “Pentagon Redirects Its Research Dollars,” New York Times, April 2, 2005, quoting officials of the Defense Advanced Research Projects Agency (DARPA) as saying that, while the amount of DARPA computer science research funding rose slightly from 2001 to 2004, the portion going to university researchers fell by about 40 percent; avail- able at http://www.nytimes.com/2005/04/02/technology/02darpa.html; accessed April 16, 2008. 14For example, the fiscal year 2008 budget request for total DOD basic research (known as 6.1 funding) declined 7.8 percent from the fiscal year 2007 budget, but total DOD uni- versity research initiatives declined more, by 14 percent. See http://www.aaas.org/spp/ rd/08ptbii4.pdf; accessed October 17, 2007. 15See “Science and Security in the Post-9/11 Environment: Foreign Students and Scholars (Updated),” available at http://www.aaas.org/spp/post911/visas/; accessed October 27, 2008. Legislation such as the USA Patriot Act of 2001 (Public Law 107-56) and the En- hanced Border Security and Visa Entry Reform Act of 2002 (Public Law 107-173) affected visa procedures.

The changing landscape: 1995-2007 49 32 percent in international student applications for the fall of 2004 com- pared with applications for the fall of 2003. Eighty percent of the schools responding reported decreases in applications for graduate engineering programs; the majority of respondents reported declines in applications from students in the two largest “sending” countries, China (76 percent of respondents reported a decline) and India (58 percent of respondents reported a decline). In the fall of 2004, the CGS reported a 6 percent decline in overall first-time international student enrollment from 2003 to 2004; engineering enrollments dropped 8 percent. This represented the third consecutive year in which the number of first-time international graduate students studying in the United States decreased between 6 percent and 10 percent from the preceding year.16 This effect appears to have mitigated by 2007, by which time enroll- ments were once again increasing. CGS survey data from 172 U.S. uni- versities (including 9 of the 10 with the largest international graduate student enrollment) on international graduate student admissions offers and enrollments for 2006 and 2007 showed that admissions offers were up 14 percent for 2006 compared with those in 2005 and that they were up 7 percent for 2007 compared with those for 2006. First-time enrollment was up 12 percent for 2006 compared with that in 2005 and was up 4 percent for 2007 compared with that in 2006. Admissions and enrollments from China and India showed the greatest increases, with engineering being the field of study showing the largest increases.17 Findings for international students overall (undergraduate and grad- uate) were reported by the Institute for International Education (IIE), which publishes the annual Open Doors: Report on International Education Exchange with support from the U.S. Department of State. According to the IIE, the total number of international students enrolled in colleges and universities in the United States increased by 3 percent over that of the previous year to a total of 582,984 in the 2006/2007 academic year; this is the first significant increase since 2001/2002. Engineering students 16See Council of Graduate Schools, “Council of Graduate Schools Finds Widespread Declines in International Graduate Student Applications to U.S. Graduate Schools for Fall 2004” and “Council of Graduate Schools Finds Decline in New International Graduate Student Enrollment for the Third Consecutive Year,” Washington, D.C., March 2, 2004, and November 4, 2004, respectively. Research reports and summaries from the CGS are available at http://www.cgsnet.org/Default.aspx?tabid=172; accessed December 11, 2007. 17See Council of Graduate Schools, “Findings from the 2007 CGS International Graduate Admissions Survey Phase III: Final Offers of Admission and Enrollment,” Washington, D.C., November 2007, available at http://www.cgsnet.org/portals/0/pdf/R_intlenrl07_III.pdf; accessed December 11, 2007.

50 assessing the impacts of changes in the it R&D ecosystem represented about 14 percent of the 2006/2007 international enrollment, up 1.5 percent from the previous year.18 However, despite the overall increase in international student enroll- ment and in engineering enrollment by international students, the IIE data showed a continuing decrease in international student enrollments in computer and information sciences. These dropped by about 40 percent from 2003/2004 through 2006/2007. In 2003/2004, about 10 percent of international students (57,739 students) were enrolled in computer and information sciences; by 2006/2007 this had fallen to 5.7 percent (33,437 students).19 Because computer science departments rely heavily on for- eign graduate students, this decrease can have a large impact on computer science degree production.20 With respect to graduate education, the Computing Research Asso- ciation’s (CRA’s) analysis of National Science Foundation (NSF) data on first-time, full-time graduate student enrollments in computer science showed a large drop in foreign student enrollments, from 6,500 students in 2001 to about 4,300 students in 2003. There was a small decline to about 4,000 students in 2004, then a small rise in 2005 to just over 4,500 stu- dents. Throughout this period, the number of foreign graduate students exceeded the number of U.S. graduate students in computer science: U.S. student enrollments rose from about 2,500 in 2001 to about 4,000 in 2003 and then declined to about 3,500 in 2005.21 Attracting talented, foreign-born students and retaining them after they graduate are important goals for enabling continued technology entrepreneurship, business formation, and job creation. As noted in Chapter 1, for at least 25 percent of U.S. engineering and technology companies started between 1995 and 2005, at least one key founder 18See Institute for International Education, 2007 data tables and summaries, available at http://opendoors.iienetwork.org/?p=113743; accessed December 11, 2007. See Institute for International Education, Open Doors 2007: Report on International Education Exchange, New York, N.Y., 2008. 19See the IIE Open Doors report statistics by field of study tabulated in Jay Vegso, “Contin- ued Drop in Foreign Total Enrollment in CIS,” CRA Bulletin, November 12, 2007, available at http://www.cra.org/wp/index.php?p=130; accessed January 2, 2008. 20In 2004, over half of doctoral degrees and over 40 percent of master’s degrees in the field of computer science were earned by foreign students. Fields that enjoyed growth in foreign student enrollments from 2005/2006 to 2006/2007 included intensive English language (30.0 percent increase), mathematics and statistics (12.3 percent increase), health professions (4.3 percent increase), physical and life sciences (3.4 percent increase), and business and man- agement (2.7 percent increase). See Jay Vegso, “Continued Drop in Foreign Total Enrollment in CIS,” CRA Bulletin, November 12, 2007, available at http://www.cra.org/wp/index. php?p=130; accessed January 2, 2008. 21Computing Research Association, “First-Time, Full-Time Graduate Enrollment in CS by Citizenship,” available at http://www.cra.org/wp/index.php?p=120; accessed February 20, 2008.

The changing landscape: 1995-2007 51 was born outside the United States. In 2005, these immigrant-founded companies (about 80 percent of which were in the fields of software and innovation and in manufacturing-related services) produced $52 billion in sales and employed 450,000 people.22 Yet an equally important goal will be to attract a larger share of U.S. citizens to advanced study in the field, as opportunities increase for foreign students to pursue informa- tion technology research and development in their own and other parts of the world. Financial Scandals and Bankruptcies (December 2001) Enron Corporation was a leading U.S. energy company that went spectacularly bankrupt in late 2001 after claiming revenues of $111 bil- lion in 2000. Its bankruptcy and the subsequent criminal charges brought against company executives, as well as other highly publicized failures such as that of WorldCom, led the U.S. Congress to respond with the ­ arbanes-Oxley Act of 2002, informally referred to as SOX.23 SOX estab- S lished new standards for boards, management, and accounting firms of U.S. public companies with respect to the visibility of and responsibility for the financial dealings within such companies. In the wake of the pas- sage of SOX, U.S. public companies have faced significant new require- ments for implementing and assessing internal controls over financial reporting. Section 404 in particular (pertaining to the certification of the integrity of the financial control structure of a firm) has proven dispropor- tionately costly for young IT companies relative to their limited resources, imposing new costs on venture firms that seek to pursue an IPO. 24 Various efforts have been advanced to propose modifications to SOX. These include reforms under consideration by the Securities and Exchange Commission (SEC) and other efforts to relax some of the most disproportionate aspects for IT start-ups pursuing an IPO. Whether the SEC reforms or others will go as fast or as far as members of the IT indus- try hope is uncertain.25 A secondary concern, more subtle to detect but 22Vivek Wadhwa, AnnaLee Saxenian, Ben Rissing, and Gary Gereffi, “America’s New Im- migrant Entrepreneurs,” Duke Science, Innovation, and Technology Paper No. 23, January 4, 2007, available at http://ssrn.com/abstract=990152; accessed December 26, 2007. 23The official name of the Sarbanes-Oxley Act of 2002 (Public Law 107-204, 116 Stat. 745) is the Public Company Accounting Reform and Investor Protection Act of 2002. 24Section 404 of SOX requires company management and an external auditor to report on the adequacy of the company’s internal controls on financial reporting; compliance requires extensive compliance documentation and testing of financial systems and controls. For a summary of a survey on the costs of SOX compliance, see http://fei.mediaroom.com/index. php?s=43&item=204; accessed May 1, 2008. 25See Sean Wolfe, “Sarbanes-Oxley Lite,” Red Herring, January 10, 2007.

52 assessing the impacts of changes in the it R&D ecosystem with potentially deeper consequences over time, is that over time, boards of directors and the corporate culture that they inspire in these young, small IT companies may shift their primary emphasis on innovation and entrepreneurship to one of regulatory compliance.26 Surviving After the Bubble Burst (2001-2004) Following the bursting of the investment bubble, the eruption of financial scandals, and the spectacular bankruptcies, firms’ focus turned both to cost cutting and to regulatory compliance. For fledgling IT start- ups, the most urgent issues involved survival and prospects for going public. For larger firms, cash conservation became far more important. IT budgets gave priority to compliance projects such as fraud detection, internal controls, risk assessment, regulations, and conforming with legis- lation on corporate governance.27 Boards became concerned with personal liability and saw a significant increase in their board duties. For many segments of the IT industry, this was the first period of prolonged spend- ing cuts. (By contrast, firms offering compliance systems faced increas- ing demand.) Spending by telecommunications carriers on equipment in the United States dropped sharply between 2000 and 2003 (from some $52 billion to $20 billion) and has only slowly increased through 2006 (to just over $24 billion).28 Similarly, growth in data-center equipment such as high-end servers stalled and moved into negative territory. Not all IT segments shrank, and IT spending as a whole grew, albeit at a far reduced pace and according to a substantially altered spending portfolio allocation (for example, companies began to spend more on compliance and control systems to help them meet regulatory requirements, sometimes at the expense of R&D and other longer-term investments). 26See Tom Perkins, “The ‘Compliance’ Board,” Wall Street Journal, March 2, 2007. A former board member of the Hewlett-Packard Company, Tom Perkins advocated the “guidance board” over the “compliance board” in this op-ed piece. 27In addition to SOX, this legislation includes the Bank Secrecy Act/Anti-Money Launder- ing Laws. The Currency and Foreign Transactions Reporting Act, 31 U.S.C. Sections 5311-5330 and 12 U.S.C. Sections 1818(s), 1829(b), and 1951-1959, also known as the Bank Secrecy Act (BSA), and its implementing regulation, 31 CFR 103, constitute a tool that the U.S. govern- ment uses to fight drug trafficking, money laundering, and other crimes. Other laws also provide tools to prevent money laundering. See Bank Secrecy Act/Anti-Money Laundering: Comptroller’s Handbook, September 2000, available at http://www.occ.treas.gov/handbook/ bsa.pdf; accessed October 19, 2007. 28“Telecommunications Industry Association 2007 Industry Playbook,” p. 4, available at http://www.tiaonline.org/gov_affairs/policyplaybook2007.swf?/policy/policyplay book2007.swf; accessed March 7, 2008.

The changing landscape: 1995-2007 53 The emphasis on cost reduction over growth investments fueled inter- est in offshoring and outsourcing.29 For the first time, even young start-up companies in Silicon Valley had to consider these strategies, despite their lack of the infrastructure that larger companies had for managing offshore functions such as R&D. The Recovery (2005-2007) The shocks to the IT R&D ecosystem eventually gave way to gradual recovery, beginning in 2005 and continuing through 2007.30 This recov- ery has positive attributes. IT budgets are growing globally, consistent with the generally positive economic climate. The enterprise market is no longer the main driver of IT innovation. The consumer market had certainly been important during the bubble period, but during the recov- ery consumers often became the dominant force, driving IT advances in many segments such as multimedia, social networking, gaming, cellular telephones, personal computers, and even automobiles. Venture capital funding was on the rise through 2007. Yet valuations remain modest, and the concern for cash conservation and sound business models remains strong. Good start-up companies generally reach successful milestones for merger and acquisition (M&A) or IPO. At the same time, there are important differences between today’s environment and the rather unique period of a decade ago. The sources of innovation are more diverse. They include the United States and other large markets in Europe and Asia. They include consumers in addition to suppliers and enterprise customers.31 The importance of consumer mar- kets has continued to grow, especially in new or growing segments such as multimedia, social networking, games, cell phones, personal comput- ers, and even automobiles. Start-up investments by venture capitalists are equally diverse. A majority of the large funds that formerly operated only in Silicon Valley now have offices in Israel, India, and China. Many of the same funds that chose to diversify their investments on the basis of 29Offshoring is the practice of moving work to developing nations. Outsourcing is the prac- tice of purchasing work that was formerly done in-house from an outside vendor. 30As this report was being prepared for publication, the effects on venture-funded and other entrepreneurial enterprises from downturns in the housing and credit markets and the economy as a whole were just beginning to be reported. See for example, Matt R ­ ichtel and Brad Stone, “Economy Has Become a Drag on Silicon Valley,” New York Times, April 9, 2008, available at http://www.nytimes.com/2008/04/09/technology/09silicon. html?ref=technology; accessed April 9, 2008. 31For a discussion of the trend toward application- and process-oriented innovation initi- ated by customers, see David Moschella, Customer-Driven IT, Harvard Business School Press, Boston, Mass., 2003.

54 assessing the impacts of changes in the it R&D ecosystem g ­ eography are also diversifying across new sectors such as “clean” tech- nology. As a result, while the total amount of venture capital raised and invested is almost identical today to what it was 10 years ago, a materially lower percentage is for U.S.-based IT investments, particularly if one takes into account dollar currency erosion. Another phenomenon associated with the 2005-2007 recovery was the rapid rise of private equity.32 Inexpensive debt instruments have made possible large recapitalizations. Even large companies may find this possi- bility appealing. In the fall of 2006, for example, Freescale Semiconductor agreed to be acquired for approximately $18 billion by a consortium of private equity firms led by the Blackstone Group. The size of this transac- tion signaled that virtually any IT company is within the reach of private equity interests, if valuation warrants it and sufficient credit is available to finance such large acquisitions. Although merger and acquisition transac- tions rose dramatically, the pace of technology IPOs has been rather slug- gish. These two trends have changed the expectations of IT entrepreneurs. It is unclear yet whether those entrepreneurs choosing to build companies for a short, independent run before acquisition, rather than with a goal of creating independent companies sustainable over the long run, will pro- duce the same kind and quality of innovation as long-run, entrepreneurial companies did in the past. It is also unclear whether the steps taken by private equity investors as they restructure the firms that they acquire will benefit users in the long run and contribute to strengthening or weaken- ing the IT R&D ecosystem.33 the Evolution of Technology Platforms The combination of hardware structures, system software, and appli- cations software that together deliver an important foundational set of IT capabilities is often called a platform. Important examples today include Web 2.0 capabilities that deliver today’s interactive Web sites, the Win- dows family of operating systems, the Intel x86 instruction set (with implementations also available from Advanced Micro Devices [AMD] and other vendors), and the combination of open-source software (Linux, the Apache Web server, MySQL, and PHP-Perl-Python) used to run dynamic Web sites and commonly known as LAMP, for the four software compo- nents that contribute to the platform. These platforms have been a critical area of IT innovation over the decades. Box 3.1 describes the evolution of major computing platforms up 32Dana Cimilluca, “Private Equity Fuels Record Merger Run,” Wall Street Journal, July 2, 2007. 33Ben Worthen, “Is Private Equity Good for Tech Users?” Wall Street Journal, June 25, 2007.

The changing landscape: 1995-2007 55 to the mid-1990s. The rest of this section describes the major IT platforms of the past decade. Notably, Web services and open-source community development have changed the fundamental nature of software, while wireless network connectivity, mobility and portability, and the emer- gence of power as a critical resource to manage have significantly affected the prevailing hardware designs. Box 3.1 The Evolution of Information Technology Platforms (1960s to Mid-1990s) •  960s and 1970s. Mainframe computers and their software systems dominat- 1 ed. A platform shift from batch to time-sharing occurred, driven by new applica- tions such as online airline reservations, while machine resources were shared across a larger user community. Technology advances made it economical to provide scaled-down “mini-computers” to work groups such as engineering teams for whom dedicated access could be justified. •  970s and 1980s. The microprocessor yielded a shift to personal computers 1 (PCs) and engineering workstations, delivering functionality and performance at a price that could be justified for an individual. The single-user nature of these machines affected the dominant operating systems, such as Disk Op- erating System (DOS) and UNIX, as well as the kinds of user applications (for example, word processing, spreadsheets, circuit design, and mechanical design). Ethernet advanced the client-server model, in which users’ computers communicated to back-end “servers” for storage, mail, and printing services on a local area network (LAN). •  id-1980s and early 1990s. Multiprocessor systems emerged for high-end M transaction processing and fault tolerance. These were often used for the server side of high performance database systems. At the client-machine level, “WIMP” interfaces—Windows, Icons, Menu, Pointing Device—became per- vasive.1 Routers and switches extended LAN technology and enabled the expansion of the National Science Foundation Network (NSFnet). Enterprise software, first in the form of relational database systems and later in the form of software tailored to perform particular enterprise functions and industry sec- tors, emerged as major IT platform elements. PC software also evolved, with programs for graphical design and page layout. •  id-1990s. The major platform shift was the emergence of the commercial M Internet. This development was driven by the underlying network equipment as well as by new end-user functionality delivered to client-side Web browsers from server-side Web servers, using the protocol architecture of the World Wide Web. New services rapidly emerged for Web directories, Internet search, e-commerce, and auctions. 1See, for example, Thierry Bardini, Bootstrapping: Douglas Engelbart, Coevolution, and the Origins of Personal Computing, Stanford University Press, Palo Alto, Calif., 2000.

56 assessing the impacts of changes in the it R&D ecosystem As platforms evolve, the established IT leaders are presented with both new opportunities and new challenges.34 Of particular note is that leadership in the definition of new platforms has implications not only for the firms that participate in that definition but for the wider IT sector. IT products and services generally become commoditized over time as mul- tiple firms acquire the know-how to supply similar, competing products; such competition has benefits in terms of lower prices for goods and ser- vices. Pressures from lower costs overseas for labor and other essentials thus require that to maintain leadership—or even a strong position—in IT, U.S. firms must constantly focus on achieving high-value innovation as a foundation for developing noncommodity products and services. 35 Basic research support can play an important role in enabling plat- form leadership. For example, NSF support for work at the National Center for Supercomputing Applications (NCSA) at the University of Illinois provided the United States with an early lead in Web browser and server technologies, even though the initial Web implementation was at a European research laboratory. Baseline: Web 1.0 Platform The year 1995 brought together the World Wide Web (WWW), the Mosaic Web browser, and the commercialized Internet. From this conflu- ence flowed the first generation of Web applications.36 The so-called Web 1.0 consists of the WWW protocol stack for the exchange of Web pages between servers and browsers, and the first generation of Web sites. Pages are described in Hypertext Markup Language (HTML) and transported between servers and clients by Hypertext Transport Protocol (HTTP). HTTP is constructed on top of the Transmission Control Protocol/Internet Protocol (TCP/IP) protocol stack originally developed for the Advanced Research Projects Agency Network, or Arpanet. Pages are identified by way of a Uniform Resource Locator (URL). Despite their genesis in university research laboratories, the Web 1.0 services quickly shifted from those for researchers and scientists to those 34For an analysis of approaches to building platform leadership including those used by Google and Qualcomm, see Annabelle Gawer and Michael A. Cusumano, “How Compa- nies Become Platform Leaders,” Sloan Management Review 49(2):28-35, Winter 2008. See also Timothy Bresnahan and Shane Greenstein, “Technological Competition and the Structure of the Computer Industry,” Journal of Industrial Economics 47(1):1-40, March 1999. 35See National Research Council, Renewing U.S. Telecommunications Research, The National Academies Press, Washington, D.C., 2006, p. 58, for a discussion of commoditization and its effects in the telecommunications sector. 36See “10 Years That Changed the World,” available at http://www.wired.com/wired/ archive/13.08/intro.html; accessed October 29, 2008.

The changing landscape: 1995-2007 57 BOX 3.2 Google: An Example of Growing from Research to Global Brand, Building on Scalable Infrastructure In 1998, Google handled 10,000 search queries per day from a “server farm” located in the dormitory room of Larry Page, computer science graduate student at Stanford University. Today, Google has 15,000 employees, diverse products, annual revenues of $15 billion, a market capitalization of more than $150 billion, and is a verb. Google’s story illustrates the critical nature of university research for start-ups and the huge difference that individuals make in the trajectory of a start-up. Larry Page and his Google cofounder Sergey Brin were research assistants at Stanford contributing to the National Science Foundation’s Digital Library Ini- tiative. Search was a natural component of this effort. Web search was not new. But Page and Brin had a new idea for improving search quality: the PageRank algorithm that weights Web page importance by the number and importance of other Web pages that link to it. Google was incorporated in late 1998 when Page and Brin received a $100,000 investment from Sun Microsystems cofounder Andy Bechtolsheim. Bechtolsheim had once been a Stanford graduate student as well: he designed the Sun Workstation under the supervision of Professor Forest Baskett, supported by the Defense Advanced Research Projects Agency (DARPA) VLSI Project. (It ran the University of California, Berkeley, UNIX operating system, engineered by Berkeley computer science graduate student and Sun Microsystems cofounder Bill Joy under the supervision of Professors Domenico Ferrari and Bob Fabry, also supported by DARPA.) Page and Brin were introduced to Bechtolsheim by Stan- ford computer science Professor David Cheriton, who had previously cofounded Granite Networks (high performance networks) and Kealia (high performance servers) with Bechtolsheim. The PageRank patent is held by Stanford and licensed to Google. focused on consumers. Web “properties” (see below) are the Web sites that are popular among the growing user community in terms of average visits by visitor, number of unique visitors, and other metrics (see Box 3.2 regarding the transition of Google from a government-funded research project to a multibillion-dollar-per-year Web property). Web Browser, Web Server, and Portals The historical role of university developers and researchers in creating Web technologies, services, and applications is remarkable. The Mosaic Web browser developed at NCSA was commercialized as Netscape Navi- gator in late 1994, achieving a memorable initial public offering in August

58 assessing the impacts of changes in the it R&D ecosystem 1995.37 The browser was freely available for evaluation, although com- mercial users needed a for-fee license. Navigator pioneered support for dynamically rendered Web pages, providing extensions to HTML and HTTP that provided useful new functionality (albeit sometimes ahead of formally adopted standards). Navigator’s market success triggered a response from Microsoft: Internet Explorer. Starting with the same (publicly available) code base developed at NCSA, Microsoft developed a sequence of versions that ultimately surpassed Navigator in technical sophistication and reliabil- ity—especially for Web pages with dynamic content and complicated rendering. Again, the business model was free software for consumers and licenses for commercial use. By 1998, Netscape surrendered in the “browser wars” marketplace competition with Internet Explorer, releasing its code as open source. This in turn gave birth to the Mozilla (later Firefox) browser and its own post-2004 competition with Internet Explorer. NCSA also developed the early Web server Hyper Text Transfer Protocol Daemon (HTTPd), which Netscape likewise commercialized. Starting with the same code base, a group independently developed the Apache Web server as open source in 1995.38 A widely used Web appli- cations stack is based on the open-source operating system Linux, the Apache Web server, MySQL (open-source data management), and PHP- Perl-Python (a scripting language for program-driven dynamic Web page behavior), known as LAMP. Netscape attempted to transition from being a software company to being a “content” company by becoming a portal—a collection of Web- based services such as user forums, e-mail, shopping services, news, Web directories, Internet search, messaging, and, more recently, Web logs (blogs) and telephony services (Voice over Internet Protocol). Popular Web sites typically provide a portal front end to back-end Web services. Directories and Search With the growth in WWW pages, finding information became a major need. “Jerry’s Guide to the World Wide Web,” the hierarchical directory 37See Audris Mockus, Roy T. Fielding, and James D. Herbsleb, “Two Case Studies of Open Source Software Development: Apache and Mozilla,” in Perspectives on Free and Open Software, Joseph Feller, Brian Fitzgerald, Scott A. Hissam, and Karim R. Lakhani, eds., MIT Press, Cambridge, Mass., 2005. 38For a history of Apache, see Audris Mockus, Roy T. Fielding, and James Herbsleb, “A Case Study of Open Source Software Development: The Apache Server,” Proceedings of the 22nd International Conference on Software Engineering, Association for Computing Machinery, 2000, pp. 263-272, available at http://opensource.mit.edu/papers/mockusapache.pdf; ac- cessed December 27, 2007.

The changing landscape: 1995-2007 59 originally behind Yahoo!, was first compiled in 1994 by two Stanford University graduate students, Jerry Yang and David Filo. A Web directory is a page with organized links to other pages. Editors typically create such directories, or authors can create entries themselves. Alternatively an automatic crawl of the Web can construct a directory. Web crawlers follow links within found pages to find subsequently reachable pages. These are analyzed for words and phrases that were used to locate the page. Search engines use a Web index to identify pages that match a particular user’s search criterion. Early Web crawlers, circa 1993, included Wanderer (developed at the Massachusetts Institute of Technology) and Aliweb (developed at the European Organization for Nuclear Research), both of which constructed limited indexes. WebCrawler (developed at the University of Washington in 1994) is considered the first to offer full-text search.39 Infoseek and Lycos (developed at Carnegie Mellon University) soon followed. Info- seek supported more complex Boolean searches. It was acquired by the Walt Disney Company in 1998 and was used by Go.com.40 Lycos, which started as a Web search site, evolved into Terra Lycos, an advertising- supported Web portal. The next wave of Web crawlers included AltaVista (from Digital Equipment Corporation [DEC]) and Excite (created by students at Stan- ford University). AltaVista was notable for the way that it implemented fast Web crawling and used multiprocessor hardware to handle the grow- ing scale of search. Excite combined search services from Magellan and WebCrawler, and was the search back end for Netscape, Apple, and Micro- soft. Excite also moved into portal services, ultimately being acquired by the ISP@Home to form Excite@Home in 1999. The Web searchers Dogpile, Inktomi (developed at the University of California, Berkeley), and Ask Jeeves emerged in 1996. Dogpile was a metasearcher, combining the results of other search engines. Inktomi used a cluster-based server architecture to improve search quality while maintaining high throughput. Ask Jeeves (now Ask.com) focused on an easy-to-use natural language system. Google (developed at Stanford University in 1998) also exploits under- lying cluster computing technology to achieve scale in processing. The system’s PageRank algorithm ordered matching pages by their impor- tance, defined as a function of how many other pages refer to the page 39WebCrawler was the search engine acquired by America Online (AOL) in 1995, in turn by Excite in 1997, and finally by InfoSpace in the wake of Excite’s bankruptcy in 2001. 40To illustrate the complex heritage of Web search products, Infoseek’s enterprise search product was sold to Inktomi Corporation in 2000, which in turn was sold to Verity in 2002 prior to Inktomi’s acquisition by Yahoo!

60 assessing the impacts of changes in the it R&D ecosystem and their importance in turn. While the search engine Overture pioneered the technology of ad placement and the revenue model of advertising- supported Web search, Google’s popularity turned ad placement into a lucrative revenue source. Ad placement is a formidable technology challenge, requiring in response to a search term that large numbers of advertisers bid—in real time—for placing their link on a user’s results page. It represents a vibrant research area at the confluence of information technology and economics. Yahoo! has assembled its current search service from Inktomi and Overture. MSN Search depended on other providers for search, particu- larly Inktomi until 2004, when it switched to its own service MSN Search. Windows Live Search débuted in 2006. The Emergence of Web Properties Beyond search and portals, Web 1.0 led to further categories of Web services. E-commerce sites such as eBay and Amazon.com are the most iconic, but there are many, many more. In September 1995, eBay started as an electronic auction site named AuctionWeb. An eBay innovation is its reputation system: sellers and buyers rate each other at the end of the auction. To simplify the process of paying for auctions, eBay acquired PayPal, a Web service that plays the role of a financial intermediary able to maintain the anonymity of buyers and sellers. eBay collects revenues from a complex fee structure related to the nature of the item listing and the final price of the auction. Amazon.com was founded in 1994 and launched in July 1995 to sell books online. Amazon.com combines buyer information with collabora- tive filtering to suggest further products of interest to the buyer. Amazon. com also presents user reviews and allows users to rate the reviewers. Many other firms have also developed a wide array of e-commerce ser- vices and capabilities. Amazon.com is particularly interesting as it also “powers” the elec- tronic stores of other Web sites by providing interfaces to its services infrastructure. Amazon.com Web Services is constructed from underlying services for scalable virtual servers, reliable network storage, message link- ing across processing and networks, comprehensive Web site traffic data, catalogs and electronic commerce, and historical pricing information. Evolution: From Web 1.0 to the Web 2.0 Platform This subsection reviews the rise of Web 2.0, a second generation of Web-based technology, services, and applications that began to emerge in the time frame of the early 2000s.

The changing landscape: 1995-2007 61 Defining Web 2.0 Web 2.0 is characterized by community-contributed and community- managed content, existing in many forms: user-contributed postings and comments; user-produced videos, indexed by user-supplied tags and augmented by comments and ratings by other users; social networking, with community formation for communications and sharing information; and photo- and link-sharing services.41 Publishers’ reaching a community that consumes and enhances content is an enabling element. The network is the platform for delivering content to applications that run inside a Web browser. In Web 2.0, the community owns and exercises control over the site information and can use programmable frameworks to make that content dynamic. Users extend the content, labeling it, rating it, and ranking it, thus providing the “social” in social networking. Contrast this with traditional media, with controlled authorship, and proscribed user enhancement of content. Web 2.0 Platform Elements The underlying technology for Web 2.0 enables Internet-based appli- cations, accessible through the user interface provided by a Web browser, to customize pages for individual users. For example, Ajax (Asynchro- nous JavaScript and Extensible Markup Language [XML]) is a Web devel- opment framework used to create modern interactive Web applications. JavaScript adds an interpreted programmatic functionality to XML, a con- tent format-specification language. The result is dynamically rendered, content-sensitive Web pages. Other programming frameworks such as Perl, Python, Ruby on Rails, and Adobe Flex provide similar capabilities using different linguistic paradigms and platform building blocks. Each has its adherents and advantages, but the effects are similar: Move your mouse over a page-rendered map to display a floating palette of links to content about that location. The Asynchronous Web browser/Web server protocol extensions enhance responsiveness through background com- munications that obviate the need to reload an entire Web page when a user action induces a change. Google, for example, provides an Ajax application programming interface (API) to permit individual Web pages to include a search bar. The page provides a context-dependent gateway into the Google back end. 41See Tim O’Reilly, “What Is Web 2.0,” September 30, 2005, available at http://www. oreillynet.com/pub/a/oreilly/tim/news/2005/09/30/what-is-web-20.html; accessed July 3, 2007. Note that “Web 2.0” is sometimes used to refer to the introduction of greatly in- creased interactivity into Web sites.

62 assessing the impacts of changes in the it R&D ecosystem Pervasive Composition in Web 2.0 Web 2.0 applications construct new applications from existing proto- cols, services, and Web sites. For example, Really Simple Syndication (RSS) is a protocol used to implement content publication by way of “feeds” to distribute news, blogs, podcasts, and other digital media. Reader appli- cations allow users to subscribe to particular feeds, view newly avail- able content, and display selected items. As another example, mashups illustrate the power of composition within Web service architectures: a mashup is a kind of Web application that uses the existing Web services frameworks to compose a new Web site from existing sites that support the necessary access APIs. Case Study: Facebook as a Platform Facebook, a popular social networking Web site that traces its history to 2004, offers a case study of how modern Web applications exploit per- vasive composition to create a new platform for application development that gives rise to a new ecosystem. Facebook’s essential structure is the social graph, associating with each user those other users to whom that person is related as a friend or participant in a common group. Applica- tions are constructed around the social graph. For example, when a user changes his or her status (e.g., indicating “I’m in the library now,” posting new photos, and so on), the change appears in the news feed of all of the user’s friends. Growing at the rate of 100,000 new users per day, in mid-2007 Face- book opened its internal applications development platform to third par- ties.42 It allows developers’ applications to access entry points in the Facebook page and to access Facebook-managed information such as user profiles, friends, photos, and event data. When a user clicks in the appropriate area of a Facebook-rendered page, a remote server deployed and managed by the application developer is invoked to process the request, compute a response, and transmit the result back to Facebook. To incentivize developers, Facebook allows them to share in the site’s advertising revenues. Significant Trend: The Rise of Open Source Community-based development existed before 1995, although the academic and research communities were the most common users of the 42See “Facebook Developers Documentation,” available at http://wiki.developers.face book.com/index.php/Main_Page; accessed October 10, 2008.

The changing landscape: 1995-2007 63 resulting software. From 1995 on, community development accelerated greatly, and the resulting software became commonly deployed in services and systems accessible by large user communities over the Internet. What Is Open Source? Open source makes a program’s source code readily available to a developer community, with specific restrictions regarding intellectual property rights. Open-source projects may be overseen by an implementer- in-chief, a small committee of developers and/or editors, or even more democratic mechanisms. The relevant intellectual property regimes cover a wide range, from those that restrict the commercial sale of software incor- porating open source to those that allow commercialization. Such software is not necessarily free. However, the business model most commonly used is based on charging for the support, packaging, and customization of the software rather than for the software itself.43 In the context of Web-based services and applications, open-source community development is the norm. LAMP44 is an example of an application environment founded on open-source components. Implications for the Software Industry Some traditional software firms have responded to the rise of open source by embracing it. As an example, IBM now supports a form of open source for its Web-based products. IBM’s Websphere middleware archi- tecture is a set of services for Web-based applications, incorporating open- source components such as Linux, Apache Web server, and Java. In 2005, IBM released Websphere Application Server Community Edition (WAS CE) as open source. WAS CE prepackages commonly needed open-source components, providing a platform within which IBM’s other proprietary components could be added. Although the software itself is available free of charge, technical support is by fee. IBM spearheads Eclipse, a community effort to create an integration environment for software tools for Java and Web services development. Its focus is on enabling the interoperation of tools from a large vendor 43Copyrighted open-source software (and some open-source software that is in the public domain) can be licensed using a variety of mechanisms. See, for example, Andrew M. St. Laurent, Open Source and Free Software Licensing, O’Reilly Media, Sebastopol, Calif., 2004; and Lawrence Rosen, Open Source Licensing: Software Freedom and Intellectual Property Law, Prentice Hall PTR, Upper Saddle River, N.J., 2004. 44See Dale Dougherty, “LAMP: The Open Source Web Platform,” January 26, 2001, avail- able at http://www.onlamp.com/pub/a/onlamp/2001/01/25/lamp.html; accessed Octo- ber 4, 2007.

64 assessing the impacts of changes in the it R&D ecosystem community. To induce tool vendors to integrate with Eclipse, IBM pro- vides visibility into its APIs and its underlying tool integration services. Open source helps third parties verify that there are no special trapdoors or APIs that give IBM’s tools any advantage over their own. In return, IBM creates a highly functional environment for software creation, lever- aging third-party tools for Java and its own Web services model, to attract applications developers. This approach is viewed as a critical response to Microsoft’s proprietary .NET Framework.45 Significant Trend: The Emergence of Mobile and Data-Center Platforms The major hardware trends of the 1995-2007 period are the following: on the Web access side, the rise of mobile devices; on the Web services side, the concentration of back-end processing into Internet data centers. This subsection reviews some of the dominant hardware trends in both. Central Processing Units Intel x86 instruction set processors, with implementations also avail- able from AMD and other vendors, have become dominant. Introduced in the late 1970s and powering the original IBM PC, the Intel x86 has driven the information technology industry to new levels of price and performance. They are now the basis for virtually all PCs, whether server, desktop, laptop, or processor cluster. A number of variations have been introduced, some optimized for high performance (the “extreme” cat- egory—suitable for use in high-end servers), others designed for good performance at lower power (the “value” category—driven primarily by the demand of laptops for long battery life), and a third category seeking a compromise between the two (the “mainstream” category—such as for standard desktop PCs). Higher levels of performance have been achieved through higher processor clock rates made possible by shrinking semicon- ductor process feature sizes and scaled voltages. Through architectural cleverness, instruction-level parallelism makes it possible to issue more than one instruction per clock cycle if the instructions do not require the same machine resources. The challenges of cooling advanced processor chips have precipitated a fundamental shift to achieving higher perfor- mance through multicore architectures rather than faster clocks. How best to harness multicore processors to achieve higher levels of performance, 45For a more detailed discussion, see Marc Erickson and Angus McIntyre, “What Is Eclipse, and How Do I Use It?” November 1, 2001, available at http://www.ibm.com/­ developerworks/opensource/library/os-eclipse.html; accessed July 3, 2007.

The changing landscape: 1995-2007 65 particularly for desktop rather than server processing, remains a research challenge. Handheld Devices Early attempts at creating convenient handheld devices met limited success. The Apple Newton offers one such example. These tended to be proprietary devices, with limited connectivity and programmability and with narrow functionality. The first broadly successful handheld was the Palm Pilot, introduced in 1995. Keys to its success were its shirt-pocket size, long battery life, one-touch PC synchronization, intuitive user inter- face with personal organizer functionality, simplified writing recognition, instant power on, and a simple programming environment that attracted a developer community to enrich and extend the platform. Following the first “organizer”-oriented devices, Apple introduced the iPod, which became the most successful consumer-oriented digital media player; Research in Motion, its Blackberry device, initially focused on e-mail; and Palm, its Treo smartphone. Apple’s recent introduction of the iPhone is illustrative of the evolution of the handheld device from a specialized gadget to a broad-based mobile computing platform. The new handheld platforms are characterized by the following: modern operating systems supported by a diverse community of application developers; rich wireless connectivity, incorporating multiple radios with transparent roaming to enable ubiquitous network computing; full participation in consumer and business-oriented Web-based services; and being deeply embedded in consumer environments. Multiple countries contribute to this environment: China produces and consumes more mobile phones than any other country, Korea pro- vides the best and fastest broadband wireless connectivity, Europe and Japan lead the world in using consumer mobile services, while the United States leads in smartphones and business mobile services. Internet Data-Center Architecture The Internet data center has emerged as a new platform for provid- ing scaled-up processing, memory, and storage resources to support the network applications used by billions of clients. Internet data centers are building-scale computing systems, containing vast numbers of process- ing clusters and storage servers. Major Web properties and many large enterprises use Internet data centers in one form or another. The ancestor of the data center is the Web hosting facility that came into being in the late 1990s. These facilities provide Web site operators with compute cycles and data storage for rent, on which they can deploy their Web servers

66 assessing the impacts of changes in the it R&D ecosystem and applications. Resources can be shared or dedicated, with trade-offs between performance, security, reliability, and cost. Hosting facilities are purpose-built, with power and cooling sufficient to support a large num- ber of machines within the building. Most Web properties began by placing their own processing clusters within third parties’ facilities. But rapid growth, coupled to a shakeout of hosting facility operators following the bust in 2000, led to a short- age of space at hosting facilities. This led firms into designing their own building-scale computer facilities, integrating processing, storage, internal and external networking, along with integral power and cooling infra- structures. The resulting data centers typically deploy 100,000 to 1 million computers within a single facility.46 The total power budget of an Internet data center must consider the demands of the power distribution system itself and the air conditioning systems needed to transfer the heat generated by the equipment as it consumes power. Efficient utilization of data-center resources, consider- ing power as a critical resource to be managed, is critical in data-center design. Mobile Applications and Communications Platforms Users demand high performance Internet access from their mobile devices. Access is achieved either by way of packet data over the cel- lular telephone system or by access over wireless local area networks (WLANs). Third-generation cellular networks are the current state of the art. They are available throughout the United States and the rest of the world. Pre-third-generation data services provided 64 kilobit per second (kbps) circuit-switched data and 384 kbps packet-switched data. This data rate is insufficient for media-rich Web page and application delivery to mobile devices. EDGE—Enhanced Data Rates for GSM (Global System for Mobile communications) Evolution—is an interim method for higher data rates. It is widely available in the United States, achieving up to 236.8 kbps (and a maximum of 473.6 kbps if twice as many slots are used for data encoding). Evolved EDGE is a next-generation technology that uses more advanced encoding methods to increase data rates up to 1 Mbps.47 46“Down on the Server Farm,” The Economist, May 22, 2008, available at http://www. economist.com/displayStory.cfm?source=hptextfeature&story_id=11413148; accessed Oc- tober 10, 2008. 47For more information, see Wikipedia, “Enhanced Data Rates for GSM Evolution,” avail- able at http://en.wikipedia.org/wiki/Enhanced_Data_Rates_for_GSM_Evolution; accessed July 31, 2007. See also “GSM/3G Market/Technology Update: EDGE Evolution,” Global Mobile Suppliers Association, December 2007, available at http://www.gsacom.com; ac- cessed February 20, 2008.

The changing landscape: 1995-2007 67 Recent third-generation deployments include support for High Speed Downlink Packet Access (HSDPA), achieving multimegabit transmissions toward the client. Up to 3.6 Mbps peak downlink data throughput has been achieved in operational networks.48 Some operators are now deploy- ing High Speed Uplink Packet Access (HSUPA), which will also allow greater uplink speeds, into the multimegabit rate.49 The alternative is WLAN technology. The IEEE 802.11 family of stan- dards, also known as Wi-Fi, makes use of license-free spectrum bands. Inexpensive access equipment that is commonly incorporated in laptop computers and other mobile devices placed it in the hands of many users and access points. Wi-Fi can obtain up to 11 Mbps (IEEE 802.11b) or 54 Mbps (IEEE 802.11g) access speeds, but sharing spectrum with other users significantly reduces the effective bandwidth. Wi-Fi is not a replacement for cellular data services; it supports shorter distances between the base station and terminal than that of cellular data services, cannot provide high bandwidth to rapidly moving users, and does not support handoffs across access points. Nevertheless, many inexpensive base stations can be deployed to cover areas dense with pedestrian users, such as city centers and industrial and university campuses. A new generation of IEEE 802.11—IEEE 802.11n—is undergoing the process of standardization. The target performance of the system is up to 248 Mbps, with 74 Mbps typical.50 Another standard, WiMAX, refers to IEEE 802.16 and was initially intended as the specification for high-bandwidth, point-to-point wireless access between fixed devices. WiMAX is also intended as a “last mile” technology to be used as an alternative to digital subscriber line (DSL) or cable connectivity to homes and offices. Interest has emerged in mobile WiMAX (IEEE 802.16e) that will provide connectivity to small access devices on the move. Access speeds of up to 70 Mbps and distances of up to tens of miles are possible, but there is a trade-off between distance to the base station and the data rates that can be achieved. User-observed access in the range of 2 Mbps is more likely.51 48 See Wikipedia, “High-Speed Downlink Packet Access,” available at http://en.wikipedia. org/wiki/HSDPA; accessed July 31, 2007. See also Global Mobile Suppliers Association survey of HSPDA and HSUPA deployments worldwide, available at www.gsacom.com; accessed February 20, 2007. 49More information is available at Wikipedia, “List of HSUPA Networks,” available at http://en.wikipedia.org/wiki/List_of_Deployed_HSUPA_networks; accessed July 31, 2007. See also www.gsacom.com; accessed July 31, 2007. 50See Wikipedia, “IEEE 802.11,” available at http://en.wikipedia.org/wiki/IEEE_802. 11#802.11n; accessed July 31, 2007. 51Data available at Wikipedia, “WiMAX,” at http://en.wikipedia.org/wiki/WiMAX; ac- cessed July 31, 2007.

68 assessing the impacts of changes in the it R&D ecosystem Voice over Internet Protocol Voice over Internet Protocol, or VoIP (also known as IP telephony) is a method for encoding voice communications using packet-switched rather than the circuit-switched techniques that are used in conventional telephone systems. Audio is sampled, digitized, and encoded into pack- ets, which are then routed on a hop-by-hop basis to the destination, where the samples are converted back into an audio form. End devices are now powerful enough to perform this conversion processing without any specialized hardware. The Session Initiation Protocol (SIP) has been standardized to establish connections between call originators and desti- nations. Using SIP, Wi-Fi VoIP-enabled handsets can operate like mobile telephones within range of a WLAN. Cellular telephones, already or soon to be equipped with Wi-Fi access, will be able to opportunistically shift between the cellular network and WLANs, on the basis of quality and cost-of-access considerations. the Evolution of Information Technology Industry Sectors The information technology sector is composed of numerous prod- ucts and services, perhaps the most significant of which have been semi- conductors, computers, and software. This section reviews the historical evolution of these essential sectors of the information technology industry, with a focus on the developments in the 1995-2007 period. Examination of these sectors illuminates the historical sources of U.S. leadership in IT as well as the increasing role that the United States and other nations have come to play in a today’s globalized IT industry. Evolution of the Semiconductor, Computer, and Software Subsectors Let us look first at the evolution of the industry from the perspective of its three essential subsectors: the semiconductor industry, the computer industry, and the software industry. The Semiconductor Industry The transistor was invented at Bell Laboratories in 1947. The technol- ogy was widely licensed, and many electronics firms quickly diversified into the industry. These firms were mostly clustered around Boston, New York, and Los Angeles. The industry was research-intensive, with contin- ual product and process advances fueling rapid growth. Particularly in the part of Northern California soon to become “Silicon Valley,” employees from some of the original firms formed successful new entrants. The first

The changing landscape: 1995-2007 69 was Shockley Laboratories, formed by William Shockley, a co-recipient of the Nobel Prize in physics for the transistor. He hired eight talented sci- entists and engineers who in 1957 left to found Fairchild Semiconductor. Fairchild in turn suffered employee defections to new spin-offs. Many of these became industry leaders, launching the Silicon Valley cluster. Silicon Valley firms accounted for about half the output of the U.S. industry by 1980.52 U.S. firms initially dominated the industry. Not only did a U.S. firm invent the transistor, but the U.S. military was the largest consumer of semiconductors and a major funder of semiconductor R&D. Texas Instru- ments demonstrated the first integrated circuit in 1958; the same year, the U.S. Air Force incorporated semiconductors in designs for the Minuteman missile.53 Defense and space programs rapidly became major customers. By 1963 the military accounted for nearly half of device sales, financ- ing around 25 percent of semiconductor R&D.54 Subsequent growth was driven by the computer industry, also based predominantly in the United States. The first challenge to the semiconductor industry in the United States came from Japan, where the initial inroads were made by firms that pioneered transistor radios and went on to transistorize televisions and other consumer electronics products. The Japanese government negoti- ated technology licenses and protected domestic producers, later sponsor- ing industrywide projects to improve production. In the 1970s, Intel and other U.S. firms pioneered memory devices. Japanese firms excelled at their manufacture, enabling them to capture a large share of the expand- ing market.55 Later the firms faced competition from established Korean firms that also became leading producers. The latter were aided by their 52Steven Klepper, “Silicon Valley—A Chip Off the Old Detroit Bloc,” in Entrepreneurship, Growth, and Public Policy, David B. Audretsch and Robert Strom, eds., Cambridge University Press, Cambridge, England, forthcoming. See also Christophe Lecuyer, Making Silicon Valley: Innovation and the Growth of High Tech, 1930-1970, MIT Press, Cambridge, Mass., 2006. 53Texas Instruments’ Jack S. Kilby received the 2000 Nobel Prize in physics for his part in the invention of the integrated circuit. See also Semiconductor Industry Association (SIA), “SIA Interactive Timeline: 1958,” available at http://www.sia-online.org/cs/about_sia/­ history; accessed October 28, 2008. 54Richard N. Langlois and W. Edward Steimueller, “The Evolution of Competitive Advan- tage in the Worldwide Semiconductor Industry,” in Sources of Industrial Leadership, Richard R. Nelson and David C. Mowery, eds., Cambridge University Press, Cambridge, England, 1999, pp. 26-27. See also Ernst Braun and Stuart MacDonald, Revolution in Miniature, Cambridge University Press, Cambridge, England, 1978. 55Arthur L. Robinson, “Perilous Times for U.S. Microcircuit Makers,” Science 208(4444):585, 1980.

70 assessing the impacts of changes in the it R&D ecosystem recruiting of large numbers of expatriates from U.S. universities and Sili- con Valley firms.56 Subsequently, U.S. firms recaptured market share in the manufacture of semiconductors, in part owing to favorable exchange rates and growing demand for microprocessors. U.S. firms improved their own manufac- turing while outsourcing production to specialist producers, known as foundries. Since the 1960s, U.S. firms had assembled and tested devices in Asia.57 In the 1980s technological and market developments enabled such Asian firms to become manufacturing specialists,58 spurring found- ries in low-wage countries such as Taiwan. Through government efforts to license process technology and transfer it to sponsored firms, Taiwan developed a vibrant industry that today ranks fourth in the world.59 Much like Silicon Valley, its industry is composed of numerous spin-offs from incumbent firms and government efforts, concentrated in the Hsinchu Science Park. Its development was also aided by Taiwanese expatriates, many of whom had been educated in the United States and had risen up through the ranks of Silicon Valley semiconductor producers.60 The Taiwanese industry enabled the formation of firms specializing in integrated design in the United States—that is, “fabless” firms (i.e., com- panies that do not operate their own fabrication facilities and therefore can concentrate their R&D resources on design rather than on manufacturing process technologies). Fabless firms coexist with vertically integrated pro- ducers and have helped the U.S. industry maintain its prowess. U.S. firms focus on innovative product concept and design, with the manufacturing taking place in Taiwan by firms that focus on advanced chip technology and high-volume processing. The U.S. firms are concentrated in Silicon Valley and other areas with major research universities, reflecting the role that leading university research centers have played in developing new design techniques, software, and engineering talent. As of 2002, 475 fab- 56Richard N. Langlois and W. Edward Steimueller, “The Evolution of Competitive Advantage in the Worldwide Semiconductor Industry,” in Sources of Industrial Leadership, Richard R. Nelson and David C. Mowery, eds., Cambridge University Press, Cambridge, England, 1999, p. 55. 57Ernest Braun and Stuart MacDonald, Revolution in Miniature, Cambridge University Press, Cambridge, England, 1978, pp. 150-151. 58Jeffrey T. Macher and David C. Mowery, “Vertical Specialization and Industry Structure in High Technology Industries,” in Business Strategy over the Industry Life Cycle, Joel C. Baum and Anita M. McGahan, eds., Elsevier, Amsterdam, 2004, p. 331. 59John A. Matthews, “A Silicon Valley of the East: Creating Taiwan’s Semiconductor In- dustry,” California Management Review 39(4):26-54, Summer 1997; and Hongwu Sam Ouyang, “Agency Problem, Institutions, and Technology Policy: Explaining Taiwan’s Semiconductor Industry Development,” Research Policy 35(9):1314-1328, 2006. 60Hongwu Sam Ouyang, “Agency Problem, Institutions, and Technology Policy: Explain- ing Taiwan’s Semiconductor Industry Development,” Research Policy 35(9):1314-1328, 2006; AnnaLee Saxenian, The New Argonauts: Regional Advantage in a Global Economy, Harvard University Press, Cambridge, Mass., 2006.

The changing landscape: 1995-2007 71 less firms were located in the United States.61 The next closest country in number of fabless firms was Canada, with 30 such companies.62 Recently the United States has specialized in more research-intensive aspects of semiconductor production, including design and the develop- ment of experimental production facilities. Semiconductor Manufacturing Technology (SEMATECH), a nonprofit consortium, was established to pursue advanced R&D in semiconductor manufacturing.63 U.S. firms have maintained their global share of patents, as appears to be true of the patenting activities of firms in other countries as well. 64 The leading U.S. firms date to the 1950s and 1960s and have been lead- ing firms for many years; likewise, the leading firms from Japan and Europe have been leaders for many years. In contrast, the newcomers have appeared from South Korea and Taiwan, which now account for 3 of the top 10 firms in the world.65 Semiconductor applications are becoming more diverse, and customers for the most advanced semiconductors are increasingly located in Asian countries, which may well enable them to move into the more research-intensive segments of the industry, posing greater challenges to the established leaders.66 The Computer Industry As with the semiconductor industry, the computer industry began principally in the United States, with the military providing early impetus 61Qualcomm, which has origins in university research, is the first “fabless” company to become one of the top 10 chip suppliers. See Mark LaPedus, “Qualcomm Cracks Top-10 in Chip Rankings,” EE Times, August 8, 2007, available at http://www.eetimes.com/news/ semi/showArticle.jhtml?articleID=201801923; accessed February 27, 2008. 62Jeffrey T. Macher, David C. Mowery, and Alberto Di Minin, “Semiconductors,” in Na- tional Research Council, Innovation in Global Industries: U.S. Firms Competing in a New World, Jeffrey T. Macher and David C. Mowery, eds., The National Academies Press, Washington, D.C., 2008. 63Between 1987 and 1996, the U.S. government provided funding to Sematech to match industry investments. The organization subsequently was renamed International S ­ ematech, with half of its current industrial participants now international semiconductor firms. See http://www.sematech.org; accessed October 30, 2008. 64Jeffrey T. Macher, David C. Mowery, and Alberto Di Minin, “Semiconductors,” in Na- tional Research Council, Innovation in Global Industries: U.S. Firms Competing in a New World, Jeffrey T. Macher and David C. Mowery, eds., The National Academies Press, Washington, D.C., 2008. 65See Wikipedia, “Semiconductor Sales Leaders by Year: Ranking for Year 2006,” available at http://en.wikipedia.org/wiki/Worldwide_Top_20_Semiconductor_Sales_ Leaders#Ranking_for_year_2006; accessed July 25, 2007. 66Jeffrey T. Macher, David C. Mowery, and Alberto Di Minin, “Semiconductors,” in Na- tional Research Council, Innovation in Global Industries: U.S. Firms Competing in a New World, Jeffrey T. Macher and David C. Mowery, eds., The National Academies Press, Washington, D.C., 2008.

72 assessing the impacts of changes in the it R&D ecosystem as a customer and a supporter of research, much of it done in universi- ties.67 The market was unclear, and early entrants were composed of office equipment firms, electronics producers, and new entrants. IBM invested heavily in research and marketing to become the early leader. It solidified its position with the System/360, intended for broad use. The only way to compete was with compatible products, such as peripherals and software. This strategy became more feasible after IBM unbundled its software in 1970 under antitrust pressure. The Japanese government orchestrated a policy to produce products compatible with IBM’s and then to surpass IBM. A number of competitive computer firms emerged, but to some degree these were hampered by their focus on a target that would soon be undermined. In contrast, European countries unsuccessfully attempted to grow national champions to compete with IBM.68 The development of minicomputers and then microcomputers ulti- mately broke IBM’s stronghold in the realm of computers. Minicomputers appealed to sophisticated users who did not need IBM’s technical support, enabling new firms to enter the industry successfully. The most promi- nent was Digital Equipment Corporation, a spin-off of the ­Massachusetts Institute of Technology’s Lincoln Laboratory, itself a product of prior military funding. The development of the microprocessor created further opportunities; the earliest producers were new firms, such as Apple, and electronics producers. The introduction of the PC by IBM fundamentally changed the market. Its modular design, based on Intel’s microprocessor and Microsoft’s operating system, defined the so-called Wintel standard that has endured. It also allowed the flourishing of independent markets for components, including software, hard disk drives, displays, and other peripherals. U.S. firms were in the forefront of many of these specialized markets and have continued to maintain their lead in such areas as microproces- sors. Simultaneously, the modular design of the PC became driven by a low-technology design and assembly process in which firms primarily integrate innovations developed by high-technology component suppli- ers.69 Managing the supply chain thus became more important than being 67See, for example, Kenneth Flamm, Creating the Computer: Government, Industry, and High Technology, Brookings Institution Press, Washington, D.C., 1988. 68Timonthy F. Bresnahan and Franco Malerba, “Industrial Dynamics and the Evolution of Firms’ and Nations’ Competitive Capabilities in the World Computer Industry,” in Sources of Industrial Leadership, Richard R. Nelson and David C. Mowery, eds., Cambridge University Press, Cambridge, England, 1999, pp. 79-132. 69Jason Dedrick and Kenneth L. Kraemer, “Personal Computing,” in National Research Council, Innovation in Global Industries: U.S. Firms Competing in a New World, Jeffrey T. ­Macher and David C. Mowery, eds., The National Academies Press, Washington, D.C., 2008.

The changing landscape: 1995-2007 73 a technological innovator.70 Many of the U.S. PC entrants were new firms, but it was difficult for them to sustain profits. Much of the production of PCs and more recently their design have shifted to lower-cost regions, especially Taiwan, which benefited from the involvement of Taiwanese expatriates. Taiwan now has a number of branded PC and component producers with significant market shares. As of 2005, Taiwanese firms produced 85 percent of laptops sold worldwide. Much of the production of these firms now occurs in China, to take advantage of the lower costs of engineers and labor there. IBM sold its PC business to Lenovo of China. Today China is the world’s largest producer of PCs.71 U.S. firms still dominate key component industries, such as micro- processors, operating systems, graphics, and hard drives. Asian firms are leaders in displays, memory devices, power supplies, batteries, mother- boards, and optical devices. Dell, Hewlett-Packard Company (HP), Apple, and Gateway have a market share of 40 percent, but U.S. firms do little production and even hand off design to others to concentrate on product planning. In this division of labor, component-level R&D, concept design, and production planning are concentrated in the United States and Japan, applied R&D and the development of new platforms in Taiwan, and product development for mature products and nearly all production and related engineering in China. The employment of engineers is stable but not growing in the United States, whereas it is growing rapidly in Taiwan.72 The growth in PC demand in Asia and the faster adoption of broadband and mobile telephony in some countries outside the United States, particularly Asia, may further accelerate the development of the industry outside the United States. The Software Industry Software is provided both by vendors, in the form of products and services, and by users. Until the 1970s, software vendors were largely com- puter producers—and hardware-producing companies such as IBM, HP, and Sun Microsystems continue to be software producers—but the advent of the PC stimulated the provision of products by software specialists. 70G. Fields, Territories of Profit: Communications, Capitalist Development, and the Innovative Enterprises of G.F. Swift and Dell Computer, Stanford University Press, Stanford, Calif., 2004. 71Jason Dedrick and Kenneth L. Kraemer, “Personal Computing,” in National Research Council, Innovation in Global Industries: U.S. Firms Competing in a New World, Jeffrey T. ­Macher and David C. Mowery, eds., The National Academies Press, Washington, D.C., 2008. 72Ashish Arora, Chris Forman, and Jiwoong Yoon, “Software,” in National Research Coun- cil, Innovation in Global Industries: U.S. Firms Competing in a New World, Jeffrey T. Macher and David C. Mowery, eds., The National Academies Press, Washington, D.C., 2008.

74 assessing the impacts of changes in the it R&D ecosystem This subsection focuses on software producers, because software services are more difficult to measure, particularly in imports and exports. 73 The software industry has been dominated by U.S. firms. The com- puter industry was concentrated in the United States when software first became unbundled from hardware. Further, the government—the military but also the National Science Foundation—has long supported computer science research, largely at universities.74 Indeed, prior to 1965, virtually anyone who learned to “program” learned to do so as part of the Cold War era Student Achievement Guarantee in Education (SAGE) Program. Consequently, U.S. firms were well positioned to be early pro- viders of software products, ranging from operating systems, to business and consumer applications, to Internet-related software such as browsers and search engines. The economics of the industry—large up-front invest- ments in software development and minimal reproduction costs—have enabled successive generations of U.S. firms to develop entrenched posi- tions in various software products, as exemplified by Microsoft in operat- ing systems. In Japan, concentrated government efforts have failed to create a successful international industry. In contrast, Israel, Ireland, and India have developed thriving, export-oriented software industries. Each excels in different areas. India has succeeded in software services; Israel has developed niche software products, particularly in security; and Ireland has thrived mainly as a home for U.S. multinationals to localize software products for Europe. Nonetheless, Israel, Ireland, and India share impor- tant characteristics. They have an abundant supply of English-speaking, technically skilled workers. Multinationals were either important pro- ducers of products—and as such, suppliers of experienced labor and potential founders of domestic firms—or were users of software that helped stimulate domestic firms. Furthermore, each of these countries had significant expatriate technical communities in the United States that played an important role in the formation and management of their domestic firms.75 73United States Government Accountability Office (U.S. GAO), Offshoring of Services: An Overview of the Issues, GAO-06-5, Washington, D.C., November 2005; U.S. GAO, International Trade: U.S. and India Data on Offshoring Show Significant Differences, GAO 06-116, Washington, D.C., October 2005; U.S. GAO, International Trade: Current Government Data Provide Limited ­ nsight into Offshoring of Services, GAO-04-932, Washington, D.C., September 2004; and I Timothy J. Sturgeon, Services Offshoring Working Group: Final Report, Industrial Performance Center, Massachusetts Institute of Technology, September 10, 2006. 74David C. Mowery and Richard N. Langlois, “Spinning Off and Spinning On: The Federal Government Role in the Development of the U.S. Computer Software Industry,” Research Policy 25(6):947-966, 1996. 75Ashish Arora, Alfonso Gambardella, and Steven Klepper, “Organizational Capabilities and the Rise of the Software Industry in the Emerging Economies: Lessons from the His-

The changing landscape: 1995-2007 75 Public policy is credited with having played an important role in Ireland’s success.76 Prompted by a poor economy and extensive emigra- tion in the 1950s, Ireland initiated a policy of luring foreign companies through incentives, particularly to attract manufacturing and service firms involved in international trade. Israel’s success is widely viewed as reflecting the broader shift engineered by the government toward R&D- intensive industries, including hardware as well as software.77 India is the exception: there, government policy was at best not harmful, and other factors such as cost advantages were more beneficial. Furthermore, pat- ent data suggest that U.S. firms have been increasing the number of their inventions in Ireland, Israel, and India.78 These countries have developed firms that continue to grow and seed new firms, resulting in potentially greater competition for U.S. software firms. Common Patterns and Future Evolution U.S. firms were early leaders in semiconductors, computers, and soft- ware. The U.S. government, and particularly the U.S. military, played a key role in their launch and early development. Government was the largest initial buyer in each industry for many years. It also funded a sub- stantial amount of research at both companies and universities. Inevitably, government demand declined in importance as each of the industries expanded, although research support in some areas such as software, device physics, and computer architectures has persisted for 20 years or more. Yet the U.S. firms that started when government was the largest customer and the firms that emerged later have both continued to domi- nate worldwide well after the decline of government support. The more labor-intensive and less research-intensive activities have moved to lower-wage countries. Often this has been aided by efforts of tory of Some U.S. Industries,” in From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Ashish Arora and Alfonso Gambardella, eds., Oxford University Press, USA, New York, N.Y., 2005, pp. 171-206. 76See, for example, Dan Breznitz, Innovation and the State, Yale University Press, New Ha- ven, Conn., 2007; and Sean O’Rian, The Politics of High Tech Growth: Development in Network States in the Global Economy, Cambridge University Press, Cambridge, England, 2004. 77For analysis of the effects of R&D policies in Israel, see, for example, Manuel Trajtenberg, “R&D Policy in Israel: An Overview and Reassessment,” in Maryann P. Feldman and Albert N. Link, eds., Innovation Policy in the Knowledge-Based Economy, Kluwer Academic Publishers, Boston, Mass., 2001, pp. 409-454. 78The number in Israel increased from about 20 patents in 1995 to about 80 in 2004; the numbers for India and Ireland in 2004 ranged from 10 to 20 and were increasing. See Ashish Arora, Chris Forman, and Jiwoong Yoon, “Software,” in National Research Council, Innova- tion in Global Industries: U.S. Firms Competing in a New World, Jeffrey T. Macher and David C. Mowery, eds., The National Academies Press, Washington, D.C., 2008.

76 assessing the impacts of changes in the it R&D ecosystem the countries’ governments to import technology from U.S. firms. The governments in Japan and Israel have been more proactive, financing new technology developments. U.S. expatriates have brought organizational experience from the United States to help establish successful firms in the latecomer countries, reinforcing those governments’ efforts. U.S. multina- tionals have also been instrumental. Many of the latecomer countries have developed impressive firms that have moved into the leadership ranks and are in turn also seeding new domestic entrants. Inevitably latecomer countries aspire to move into more sophisticated, research-intensive activities. U.S. firms have taken advantage of these new opportunities by increasing not only development but also research in latecomer countries. At the same time, local demand for IT products in the latecomer countries is growing, and these nations are moving into the forefront in some areas as lead users. These developments portend greater competition for U.S. IT firms in the future, while simultaneously presenting new global expansion opportunities. National Clusters Firms in certain industries often form in close proximity within a country to create clusters. A cluster develops its own dynamics, which allow it to evolve. Clusters enable superior access to specialized labor, suppliers, and information. Clusters often grow up around major research universities, and a strong cluster in turn strengthens nearby institutions like universities, which provide the cluster with trained students and technical knowledge. Cluster firms in the same line of business observe one another and compete fiercely. Complementarities also develop. In the United States, for example, Silicon Valley has not only leading semicon- ductor firms, but also some of the largest equipment manufacturers and design software firms. Often, a small number of “root firms” are respon- sible for spinning off a large number of “child firms,” thereby forming clusters.79 Rapidly developing technology creates the opportunities for the creation of new firms to continue. Thus, clusters are important parts of the overall U.S. IT R&D ecosystem. IT is particularly interesting for the way that it has created entirely new industries and transformed old ones.80 It is the richness and power 79See Steven Klepper, “Employee Startups in High-Tech Industries,” Industrial and Cor- porate Change 10(3):639-674, September 2001; and Steven Klepper, “Employee Start-Ups in High Tech Industry,” in Stefano Breschi and Franco Malerba, eds., Clusters, Networks, and Innovation, Oxford University Press, USA, New York, N.Y., 2006. 80For a discussion of this phenomenon in the case of Silicon Valley, see M. Kenney and D. Patton, “The Coevolution of Technologies and Institutions: Silicon Valley as the Iconic High-

The changing landscape: 1995-2007 77 of these “tools for thought”81 that has provided so many opportuni- ties for new firms to form. The United States continues to have two of the world’s most sophisticated IT clusters, the San Francisco Bay Area (Silicon Valley) and Boston (Route 128).82 It should not be surprising that outstanding universities and a strong financial community characterize both regions. There are a number of other, smaller national IT clusters within the United States, including those in San Diego, California; Seattle, Washington; Irvine, California; Austin, Texas; and the Washington, D.C., area. Further, an extensive network of intermediaries has developed to support IT (and other) entrepreneurs.83 International Development of Clusters Since 1995, there has been an international proliferation of IT clusters that have significant R&D underway. They vary in size and with respect to the conditions that motivated their growth. In some of these, such as India (in particular, Bangalore), and China (Beijing and to a lesser degree Shanghai), wage advantages were significant factors. However, in many of these locations, wages were not key factors or, in the case of Scandina- via’s wireless technology cluster based around Ericsson and Nokia, were of no significance at all. In the cases of the IT clusters in Israel and Ireland, initially they enjoyed advantageous wages in comparison with wages in the United States and Western Europe, but for these two nations, this is no longer the case. A number of these clusters are based on narrow specializations. Taiwanese firms have developed strong niches in electronics assembly, particularly for desktop and notebook computers, as well as semicon- ductor foundries for chip fabrication. Taiwan has parlayed its strengths Technology Cluster,” in P. Braunerhjelm and M. Feldman, eds., Cluster Genesis: Technology- Based Industrial Development, Oxford University Press, Oxford, England 2006, pp. 38-60. 81Howard Rheingold, Tools for Thought: The People and Ideas of the Next Computer Revolution, Simon and Schuster, New York, 1985; and Stephen S. Cohen, John Zysman, and Bradford J. DeLong, “Tools for Thought: What Is New and Important About the ‘E-conomy’?” January 1, 2000, available at http://repositories.cdlib.org/brie/BRIEWP138/; accessed July 3, 2007. 82A key book on the early development of Silicon Valley is Christophe Lecuyer’s ­Making Silicon Valley: Innovation and the Growth of High Tech, 1930-1970, MIT Press, Cambridge, Mass., 2006. For a variety of perspectives on the formation of Silicon Valley, see M. Kenney, ed., Understanding Silicon Valley: Anatomy of an Entrepreneurial Region, Stanford University Press, Stanford, Calif., 2000. See also E.M. Rogers and J.K. Larsen, Silicon Valley Fever, Basic Books, New York, 1984; and AnnaLee Saxenian, Regional Advantage: Culture and Competition in Sili- con Valley and Route 128, Harvard University Press, Cambridge, Mass., 1994. 83For information on entrepreneurial support networks, see M. Kenney and D. Patton, “Entrepreneurial Geographies: Support Networks in Three High-Tech Industries,” Economic Geography 81(2):201-228, 2005.

78 assessing the impacts of changes in the it R&D ecosystem into higher-volume consumer products, such as cellular telephones, per- sonal digital assistants (PDAs), and MPEG-1 Audio Layer 3 (MP3) play- ers, although the actual assembly is increasingly being shifted to China. Despite the gradual increase in Taiwanese semiconductor design firms, the United States still dominates in design. These smaller nations have found valuable niches, yet because of their size there is little concern that they will dominant the global indus- try. India and China are different in that they have wages significantly lower than those in the developed nations. In population and gross domestic product (GDP), they dwarf Ireland, Israel, and Scandinavia. Further, given their rapid growth they are increasingly significant in the global economy. Table 3.1 documents the relative size and rate of growth of the developing IT industry in India, Ireland, and Israel. China is not included in Table 3.1 because it has limited software exports at the pres- ent time.84 India Understanding India’s role in the global IT economy is difficult because of the speed with which it is changing and the fact that even today most of the work is development, not cutting-edge basic research. A decade ago, IT R&D in India was confined to the small operations of a few U.S. pioneers such as HP, Motorola, and Texas Instruments. India did not begin as a performer of development; rather it entered the IT economy by providing programmers and doing routine programming work. Today, the vast majority of Indian software professionals continue to do such routine work. What has changed is that many European and U.S. multinational corporations have established research and develop- ment facilities in India. So India, which only 5 short years ago would have hosted very little R&D, is becoming an increasingly significant IT R&D center largely because of the decisions by U.S. IT firms to take advantage of an increasingly rich IT ecosystem. The significance of the changes in India is worth some attention. India’s entry into the global industry dates to 1974, when Burroughs Corporation asked Tata Consultancy Services to supply programmers to install system software for its U.S. client.85 In the 1980s, a few U.S. firms set up facilities in Bangalore where Indian engineers and programmers 84Association for Computing Machinery Job Migration Task Force, Globalization and Off- shoring of Software: A Report of the ACM Job Migration Task Force, W. Aspray, F. Mayadas, and M. Vardi, eds., Association for Computing Machinery, New York, N.Y., 2006, pp. 120-121. 85S. Ramadori, Chief Executive Officer, Tata Consultancy Services, personal interview by Rafiq Dossani, Walter H. Shorenstein Asia-Pacific Research Center, Stanford University, 2002.

The changing landscape: 1995-2007 79 TABLE 3.1  Software Exports from India, Ireland, and Israel, Selected Years from 1990 to 2005 (in $ millions, except where otherwise noted) India Ireland Israel 1990 105 2,132 90 2000 6,200 8,865 2,600 2002 7,500 12,192 3,000 2003 8,600 11,819 3,000 2005 17,100 18,631 3,000 Number employed (2003) 260,000 23,930 15,000 Revenue/employee (2003) 33,076 493,988a 273,000 Number employed (2005) 513,000 24,000 NA Revenue/employee (2005) 33,333 776,000a NA NOTE: Data for India are from R. Heeks, India’s Software Industry: State Policy, Liberalisation and Industrial Development, Sage Publications, New Delhi, India, 1996; and Nasscom, Review of the Indian IT Industry, New Delhi, India, Nasscom, 2003-2006. Data for Ireland are from http://www.nsd.ie/htm/ssii/stat.htm, downloaded September 26, 2006. Data for Israel are from http://www.iash.org.il/Content/SoftwareInds/SoftwareInds.asp, downloaded Au- gust 31, 2003, and http://www.israel21c.org/bin/en.jsp?enDispWho=InThePress&enPage =BlankPage&enDisplay=view&enDispWhat=Zone&enZone=InThePress&Date=08/11/05, downloaded September 26, 2006. Data for Ireland prior to 2003 are in euros (converted at 1 euro = $1.043, the rate on January 5, 2003). From 2003 on, data are converted at 1 euro to $1.26, the rate in January 2004. The most recent figures for Israel are for 2001.   aSands (p. 45) argues that the revenue/employee for Ireland is overstated because of in-country transfers and should be about $160,000. If so, total exports in this table are over- stated by a factor of three. See A. Sands, “The Indian Software Industry,” in A. Arora and A. Gambardella, eds., From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Oxford University Press, New York, N.Y., 2005, pp. 41-71. This finding is seconded by Dan Breznitz, who calculated that sales per employee for Irish-owned software firms hit a high of $104,000 per employee in 1999. See Dan Breznitz, Innovation and the State, Yale University Press, New Haven, Conn., 2007. As outlined in the text of this chapter, a number of countries have developed significant IT sectors. Two of them, India and China, are, by virtue of their size, competiveness, and close links to the U.S. IT sector, of the greatest significance for the U.S. IT R&D ecosystem. Source: Rafiq Dossani and Martin Kenney, “The Implications of Globalization for Software Engineering,” in National Academy of Engineering, The Offshoring of Engineering: Facts, Un- knowns, and Potential Implications, The National Academies Press, Washington, D.C., 2008. Adapted with permission of the National Academy of Engineering, 2008. developed products for global and domestic markets. Knowledge of the capabilities of Indian programmers and engineers gradually spread, but even in 1990, the Indian IT industry was not well known. By 1995, Indian IT industry growth had quickened, and at that time there were 27,500 Indians in the software services export industry. By 2006, this number had increased to 706,000. India has the second-largest number of soft- ware and software services industry workers in the world, following the

80 assessing the impacts of changes in the it R&D ecosystem United States. The 2007 pace of growth is not slackening, although there are indications of labor shortages by 2010. After 2000, India’s domestic market grew rapidly, from a GDP of $460.2 billion to $805.7 billion in 2005.86 By early 2007, India was adding 5 million mobile phone users per month (although subscribers pay only $8 per month on average).87 Similarly, broadband usage in business and among the higher-income groups in India is growing. Even though India is a moderate-sized market, if current growth rates continue, it will be one of the five largest in the world. There are indications that as respect for their capabilities increases, Indian firms are becoming trusted advisers for Global Fortune 1000 firms. There is some indication that Indian firms are surpassing the U.S. firms on software quality metrics and might be overtaking U.S. providers on quality as well as labor cost advantage.88 U.S. rivals have expanded their Indian operations. IBM now has approximately 60,000 employees in India and Electronic Data Systems (EDS) approximately 35 percent of its global total. India is likely to be a major recipient of further offshoring. In software itself, India is also developing capabilities. According to Google’s official blog, for example, Google Finance “started as a small project led by a few engineers in Bangalore and later joined by more engineers and finance enthusiasts in Mountain View and New York.”89 All major U.S. software and Internet firms now have large and growing operations in India. Many employ more engineers in India than in any other nation outside the United States. Semiconductor Design  India is also becoming a major semiconductor design center. Indian firms are part of the global value chain for integrated circuit (IC) design, having moved from verification, physical design, and silicon production engineering to higher-value work in architecture and design of analog and digital circuits.90 Indian firms are now marketing 86World Bank, “Key Development Data and Statistics,” 2007, available at http://devdata. worldbank.org/external/CPProfile.asp?PTYPE=CP&CCODE=IND; accessed June 20, 2007. 87Ruth David, “Vodafone Wins Stake in India Cell Phone Market,” Forbes, February 2007, available at http://www.forbes.com/business/2007/02/12/essar-hutch-vodafone-cx_rd_ 0212bid-update2.html; accessed June 20, 2007. 88Leonard Lynn and Harold Salzman, “The ‘New’ Globalization of Engineering: How the Offshoring of Advanced Engineering Affects Competitiveness and Development,” paper presented at the Sloan Industry Studies Annual Meeting, Boston, Mass., 2007. 89Google, “Spring Is the Season for Love (and Data),” posted on The Official Google Blog, March 21, 2006, available at http://googleblog.blogspot.com/2006/03/spring-is-season-for- love-and-data.html; accessed June 20, 2007. 90Rafiq Dossani and Martin Kenney, “Implications of Globalization for Software Engi­ neering,” in National Academy of Engineering, The Offshoring of Engineering: Facts,

The changing landscape: 1995-2007 81 their expertise to provide end-to-end solutions. They can manage the design handoff to IC producers such as Taiwan Semiconductor Manufac- turing Company (TSMC) directly, without involving the end customer. Many U.S. semiconductor firms now have engineering operations in India,91 often their largest outside the United States.92 Members of senior management of these operations often have degrees and experience in the United States.93 The number of Indian very-large-scale integrated circuit (VLSI) design engineers was 11,300 in 200594 and was projected to grow to 33,135 by 2010. For the year 2005, revenues were estimated to be $583 million; they are expected to reach $2 billion by 2010.95 The leading semiconductor design software firms are also increasing their presence in India, both to service the local market and to support the global market. Although there are currently no commercial fabrication facilities in India, the design functions for many parts of the value chain in semiconduc- tor, design software, and equipment suppliers are beginning to emerge. There is sufficient anecdotal evidence to suggest that India is quite rapidly becoming a force in semiconductor design. Computer Networking Equipment  Networking equipment manufactur- ing has undergone a severe shakeout since the collapse of the IT bubble. Shifting development to low-cost regions is certainly a consideration. The leading firms all have engineering operations in India. As with semicon- ductors, India produces a small amount of networking equipment inter- nally or for export. Nevertheless, this situation appears to be changing. For example, Cisco Systems’ Globalization Center and chief globalization ­ n­knowns, and Potential Implications, The National Academies Press, Washington, D.C., 2008, U pp. 49-68. 91See Jeffrey T. Macher, David C. Mowery, and Alberto Di Minin, “Semiconductors,” in National Research Council, Innovation in Global Industries: U.S. Firms Competing in a New World, Jeffrey T. Macher and David C. Mowery, eds., The National Academies Press, Washington, D.C., 2008; and Clair Brown and Greg Linden, “Semiconductor Engineers in a Global Economy,” in National Academy of Engineering, The Offshoring of Engineering: Facts, Unknowns, and Potential Implications, The National Academies Press, Washington, D.C., 2008, pp. 149-178. 92It is important to note that some semiconductor firms have more employees in their manufacturing or semiconductor assembly and test operations in East Asian nations such as China and Malaysia. 93Rafiq Dossani and Martin Kenney, “Implications of Globalization for Software Engineer- ing,” in National Academy of Engineering, The Offshoring of Engineering: Facts: Unknowns, and Potential Implications, The National Academies Press, Washington, D.C., 2008, pp. 49-68. 94Indian Semiconductor Association, “Semiconductor Driven Industry in India—A Per- spective,” February 2006, PowerPoint presentation made available to committee member Martin Kenney by Rajendra Khare, president, Indian Semiconductor Association. 95The committee thanks Rajendra Khare, president of the Indian Semiconductor Associa- tion, for these data.

82 assessing the impacts of changes in the it R&D ecosystem officer are now located in India. Cisco plans to double or triple its current number of approximately 2,000 employees in India.96 Cisco already has more employees in India than it has anywhere else outside the United States. The same is true of Juniper Networks,97 and even the China-based Huawei Technologies employed about 1,300 engineers in Bangalore in late 2006.98 The anecdotal evidence suggests that the number of telecommu- nications equipment engineers in India will continue to grow, although most of this employment will be in multinational corporations (MNCs). It is likely that even with this growth, the number of telecommunica- tions equipment engineers in India will trail that of the United States and China. IT Start-ups in India  The Indian ecosystem for IT start-ups is compli- cated.99 The first category is “traditional,” India-only start-ups whose headquarters and design operations are in India—for example, the U.S.- venture-funded Tejas Networks. The second category is firms established in the United States but having operations in India. U.S. operations may be limited to a headquarters with finance and marketing functions, with the remaining operations in India; the latter has the majority of the employees and undertakes all product development. Alternatively, there are start- ups that have almost all of their employees in the United States with an engineering facility in India; Tensilica Technologies India is an example of this organization.100 Finally, there are firms that have the management and core engineering team in the United States, with the rest of the engi- neering team in India. An example is Telsima, Inc., established in 2004 with international venture funding to develop WiMAX-based broadband 96Cisco, “Cisco Provides Update on US $1.1 Billion Investment in India,” Decem- ber 6, 2006, available at http://newsroom.cisco.com/dlls/global/asiapac/news/2006/ pr_12-06b.html?CMP=AF17154&vs_f=News@Cisco:+Technology+Innovation+&+Development+ News&vs_p=News@Cisco:+Technology+Innovation+&+Development+News&vs_k=1; ac- cessed June 20, 2007. 97R. Savitha, “Juniper Networks Office in Chennai by Year-End,” Hindu Business Line, May 26, 2007, available at http://www.thehindubusinessline.com/2007/03/27/stories/ 2007032703750400.htm; accessed June 29, 2007. 98Ravi Sharma, “Huawei Keeping Fingers Crossed on Simpler Work Visa Norms for Chinese Personnel,” The Hindu, December 17, 2006, available at http://www.hindu.com/ 2006/12/17/stories/2006121701920400.htm; accessed June 20, 2007. 99Rafiq Dossani and Martin Kenney, “The Evolving Indian Offshore Services Environment: Greater Scale, Scope and Sophistication,” Sloan Industry Studies Working Papers, Num- ber WP-2007-34, 2007, available at http://www.industry.sloan.org/industrystudies/­ workingpapers/index.php; accessed October 25, 2007. 100Ashish Dixit, “Tensilica Technologies India: An Update,” paper presented at the Glo- balization of Services—The Second Annual Conference, Stanford, Calif., December 12, 2006, available at http://iis-db.stanford.edu/evnts/4587/Tensilica.pdf; accessed June 20, 2007.

The changing landscape: 1995-2007 83 wireless access and software for mobility solutions for media-rich (data- intensive) applications.101 Measuring the technical sophistication of these start-ups is difficult, but anecdotal evidence suggests that at least some of them are quite sophisticated. Examples of recent acquisitions of Indian start-ups by U.S. firms102,103 indicate that these firms are creating value and that more start- ups should be expected. China China has become the undisputed global IT equipment-manufacturing leader, which has helped fuel its rapid economic growth. China’s success in IT equipment production is different from that of other developing nations in that these are high-technology products.104 Chinese exports of IT equipment increased from $645 million in 1990 to $81 billion in 2004. 105 Nevertheless, IT manufacturing is mostly a low-margin business. Thus far, with the possible exception of firms such as Huawei and Lenovo, Chinese firms do not compete as global brands; they manufacture for others.106 The Chinese government is furthering its domestic industry while also encouraging foreign firms to produce and perform R&D locally. It has steadily increased R&D funding in engineering and the sciences and also encourages the development of its local technology standards in fields such as wireless. Some have interpreted this as “a strategy to domi- nate the global market for information technology goods.”107 Another interpretation is that the Chinese government and firms seek to decrease their dependence on foreign standards and patents, mostly held by U.S. firms, and for which they must pay royalties. The Chinese government 101Telsima, Inc. 2007. “Corporate Brochure,” available at http://www.telsima.com/pic/ pdf/download/Corporate_Brochure.pdf; accessed June 20, 2007. 102Synopsys, “Synopsys Acquires ArchPro Design Automation,” June 18, 2007, available at http://synopsys.mediaroom.com/index.php?s=43&item=468; accessed June 20, 2007. 103Computergram International, “Broadcom Acquires Indian Fabless Chip Firm Armedia,” July 6, 1999, available at http://www.highbeam.com/doc/1G1-55071676.html; accessed June 20, 2007. 104Dani Rodrick, “What’s So Special About China’s Exports?” CEPR Discussion Paper No. 5484, January 2006, available at http://ssrn.com/abstract=902348; accessed July 2, 2007. 105Jason Dedrick and Kenneth L. Kraemer, “Is Production Pulling Knowledge Work to China? A Study of the Notebook PC Industry,” Computer, July 2006, p. 37, available at http:// pcic.merage.uci.edu/papers/2006/dedrick.pdf; accessed June 22, 2007. 106Ibid. 107David Lague, “China Overtakes U.S. as Supplier of Information Technology Goods,” International Herald Tribune, December 11, 2005, available at http://www.nytimes.com/ 2005/12/11/business/worldbusiness/11cnd-hitech.html?ex=1291957200&en=748942b64ba 7f2b9&ei=5090&partner=rssuserland&emc=rss; accessed June 22, 2007.

84 assessing the impacts of changes in the it R&D ecosystem places significant pressure on foreign MNCs wishing to operate in China to establish joint ventures through which the Chinese partner can learn about foreign technologies. Semiconductors  China has become the largest single market in the world for semiconductors.108 In 2005, it accounted for 24 percent of global semi- conductor production and was responsible for 90 percent of consumption growth even as it produced only 7 percent of the worldwide total.109 As a result, China runs a significant trade imbalance in semiconductors, which is considered a serious issue by Chinese policy makers. A Chinese fabless semiconductor design industry is now emerging. The firms are small, yet most are enjoying rapid growth. In semicon- ductor fabrication China is a minor player, although with the facilities already announced or under construction in 2006, its production capac- ity could approach 10 percent of worldwide wafer production.110 Much of this foundry capacity competes directly with Taiwan, not with U.S. manufacturers, and little of it will be at the most advanced levels of technologies.111 Software and Services  The Chinese software services industry is much smaller than its Indian counterpart.112 According to the Chinese Software Industry Association, there are 300,000 workers employed in more than 6,000 firms; of these workers approximately 160,000 are software profes- sionals, about 25 per firm.113 According to the Ministry of Commerce, the revenues of the Chinese IT and software services industry increased from $7.17 billion in 2000 to $19.35 billion in 2003. During the same period, software exports increased from $250 million to $2 billion. 114 A 108PricewaterhouseCoopers, China’s Impact on the Semiconductor Industry 2006/Update, 2007, p. 7, available at http://www.pwc.com/extweb/onlineforms.nsf/weblookup/USENGTCE NChina’sImpactontheSemiconductorIndustry-2006Update-DownloadForm?opendocument; accessed June 22, 2007. 109Ibid., p. 1. 110Ibid., p. 24. 111Ibid., p. 25. 112This subsection, “Software and Services,” draws heavily on Chapter 3 in Association for Computing Machinery Job Migration Task Force, Globalization and Offshoring of Software: A Report of the ACM Job Migration Task Force, W. Aspray, F. Mayadas, and M.Y. Vardi, eds., Association for Computing Machinery, New York, N.Y., 2006. 113T. Tschang and L. Lan Xue, “The Chinese Software Industry,” in A. Arora and A. Gambardella, eds., From Underdogs to Tigers: The Rise and Growth of the Software Industry in Brazil, China, India, Ireland, and Israel, Oxford University Press, USA, New York, N.Y., 2005, pp. 131-167. 114China Software Industry Association, “China Software Export Achieved 7 Times Growth in Five Years,” 2005, available at http://www.csia.org.cn/chinese_en/index/; accessed March 2005.

The changing landscape: 1995-2007 85 recent report notes that China's total IT services revenues are rising but are barely half of India’s $12.7 billion.115 Growth is driven by internal demand, and exports make up only 10 percent of total annual software service revenues. (For comparison, the global IT services market in 2006 was $671.4 billion.116) The Chinese software export industry faces many obstacles. It is frag- mented, with few firms capable of undertaking large projects.117 As China is the world’s manufacturer, many of its products contain embedded software. Some portion of this work may be relocated to China.118 The Chinese firms providing IT services to the West are mostly small. Western firms have also established software subsidiaries in China to support their growing Chinese businesses and to provide offshore services to Asia- Pacific nations, particularly Japan.119 IT R&D in China  In 2006, the Organisation for Economic Co-operation and Development (OECD) announced that China was the world’s sec- ond-largest R&D spender.120 Although accurate figures on R&D spending are difficult to come by, there can be little doubt that R&D is growing rapidly. A number of U.S., European, and particularly Taiwanese electron- ics firms have established R&D facilities in China. It is likely that most of these facilities focus on adapting products for the local market or on production engineering, but some have global product mandates or are 115K.C. Krishnadas, “Fragmented China Software Sector No Match for India, Report Finds,” EE Times, February 18, 2005, available at http://www.informationweek.com/story/ showArticle.jhtml?articleID=60402234&tid=5979; accessed October 4, 2007. 116Robert De Souza, Kathryn Hale, Freddie Ng, and Akimasa Nakao, “Dataquest Alert: IT Services Forecast, Worldwide, 2007-2011 (Update),” G00152463, October 9, 2007, available at http://www.gartner.com/DisplayDocument?id=530210; accessed February 27, 2008. 117K.C. Krishnadas, “Fragmented China Software Sector No Match for India, Report Finds,” EE Times, February 18, 2005, available at http://www.informationweek.com/story/ showArticle.jhtml?articleID=60402234&tid=5979; accessed October 4, 2007. 118C. Brown and G. Linden, “Offshoring in the Semiconductor Industry: A Historical Perspective,” in Susan M. Collins and Lael Brainard, eds., Brookings Trade Forum, Brookings Institution Press, Washington, D.C., 2005. 119“China Becomes Japan’s Biggest Software Outsourcing Base,” Xinhua, April 12, 2007, available at http://news.xinhuanet.com/english/2007-04/12/content_5968762.htm; ac- cessed June 22, 2007. 120Organisation for Economic Co-operation and Development, “China Will Become World’s Second Highest Investor in R&D by End of 2006, Finds OECD,” 2006, available at http://www.oecd.org/document/26/0,2340,en_2649_201185_37770522_1_1_1_1,00.html; accessed June 22, 2007.

86 assessing the impacts of changes in the it R&D ecosystem doing research for their firms’ global operations.121 The most celebrated of these, the Microsoft Research Asia laboratory in Beijing, employed approximately 300 scientists in 2007.122 Furthermore, major Chinese firms such as Huawei and Lenovo are investing heavily in R&D. At the current level, China is one of the largest R&D performers in the world. At least in some cases, the R&D is already world-class. Data Communications Equipment  The Chinese data communications equipment industry has grown rapidly, with two globally recognized Chinese firms: Huawei Technologies and ZTE Corporation. Huawei sales reached $11 billion in 2006, with 65 percent from outside China.123 ZTE’s global sales reached $2.8 billion.124 Huawei’s customers include major Western operators, such as British Telecom. Nevertheless, there is no evidence that China is generating large numbers of telecommunications equipment start-ups, despite the fact that it has substantial venture capital resources and a rapidly developing internal market for telecommunica- tions products. In the future, it is possible that entrepreneurs may begin establishing start-ups. Conclusions  Chinese IT R&D will continue its rapid growth, given its past growth, the inherent commercial opportunities, and the importance given to it by the Chinese government. The enormous buildup in IT pro- ductive capacity in China will become a magnet for production engineer- ing and higher-level R&D. Given the likely growth of China’s domestic market, no major IT firm can afford to ignore the market, and it will be necessary to support that market with some domestic production. Given China’s expanding labor pool of low-cost engineers, multinational cor- porations experiencing pressure on margins are likely to expand their engineering activities there.125 In terms of R&D, China is rapidly increas- 121Xiaohong Quan, “Multinational Corporations’ R&D in China: IP Protection and Inno- vation for the Global Market,” PowerPoint presentation, November 29, 2005, available at http://iis-db.stanford.edu/evnts/4317/Xiaohong_(Iris)_Quan_presentation.pdf; accessed June 22, 2007. 122Microsoft Research, 2007, available at http://research.microsoft.com/aboutmsr/labs/ asia/; accessed June 22, 2007. 123Huawei, “Financial Highlights,” 2007, available at http://www.huawei.com/corporate_ information/financial_highlights.do; accessed June 22, 2007. 124ZTE, “Corporate Reports,” 2007, available at http://wwwen.zte.com.cn/main/files/ 2008/04/09/333826174089.pdf; accessed June 22, 2007. 125China’s current supply of “engineers” who are comparable to engineers in the U.S. workforce is likely smaller than Chinese government data might at first suggest. However, graduation rates are rising, and China is rapidly increasing its production of engineering and technology Ph.D.’s. See V. Wadhwa, G. Gereffi, B. Rissing, and R. Ong, “Where the En- gineers Are,” Issues in Science and Technology, Spring 2007. (But see also Denis Fred Simon,

The changing landscape: 1995-2007 87 ing its share of total global R&D. Some Chinese firms are already global competitors in the IT industry, and there are likely to be more. Taiwan Taiwan gradually worked its way up the value ladder from produc- ing the simplest parts and assembling consumer electronics products to designing and engineering all but the most sophisticated products. In the process, Taiwanese manufacturers have become among the largest elec- tronics firms in the world. In addition, Taiwan established a world-class semiconductor fabrication industry supporting the U.S. fabless semicon- ductor industry. Taiwanese firms have become an integral part of global commodity chains developed by U.S. firms. Facing competition from Japan, U.S. firms sought lower prices by shifting assembly and low-end manufacturing to Taiwan and Korea in the 1960s. U.S. firms and their Japanese competitors established produc- tion facilities in Taiwan, sourcing low-technology components from local vendors. The microcomputer provided an opening for Taiwanese firms to supply low-technology parts and components to PC makers. These firms soon evolved into contract assemblers, drawing on a base of smaller component makers. By the early 1990s, Taiwan had become the global center for the production of every component in the PC except for the microprocessor, dynamic random-access memory (DRAM), hard drive, graphics chips, and software. By the late 1990s, Taiwanese firms were assembling not only desktop PCs but notebook computers as well. A U.S. vendor provided the basic specifications for a desktop or notebook PC, and a Taiwanese vendor designed, assembled, and shipped the computer with the vendor’s name. Taiwanese firms began to feel margin pressure as local wages increased. They began offshoring production to China. By 2000, every significant Taiwanese electronics firm had a presence in China. Taiwan specialized in the most sophisticated manufacturing, design, mar- keting, and other headquarters functions. The assemblers have diversified from PCs into other consumer electronics products. They did this as con- tract manufacturers and not under their own brands, unlike Korea with its Samsung Electronics and LG Electronics global brands. Taiwan has also become a world leader in contract semiconduc- tor fabrication. Its foundries are enormously capital-intensive facilities, deploying advanced fabrication technologies. The availability of these foundries has enabled U.S. entrepreneurs to establish a significant number of successful specialty IC design firms. Taiwan’s own IC design sector Cong Cao, Ron Hira, and Rick Rashid, “Not Enough U.S. Engineers? (FORUM),” Issues in Science and Technology, Summer 2007.)

88 assessing the impacts of changes in the it R&D ecosystem employed approximately 14,000 chip designers in 2004 and generated $8.6 billion in revenue in 2005.126 The Taiwanese design industry faces many challenges. Taiwan’s wages are nearly double those of China and India. Taiwanese chip designers’ capabilities match those of Taiwanese assem- blers but lag behind the leading-edge system-level designers, making it difficult for them to get the most lucrative design wins. The movement of the Taiwanese assemblers to China may also relocate business opportuni- ties to lower-cost Chinese designers. To summarize, Taiwan’s position is as a supplier to global firms, pri- marily from the United States. Although the desktop or laptop PC may have been designed in Taiwan and manufactured in China, most of the added value is captured on the one hand by the U.S. brands that control the distribution channels and on the other by the U.S. firms that provide the high-value components (such as the microprocessor and the software). In fact, using these lower-cost Asian suppliers has kept the U.S. PC indus- try competitive. Taiwanese firms also provide the critical foundries for the smaller but highly profitable U.S. specialty chip design firms. The Taiwan- ese IT industry has evolved a mutually beneficial division of labor with its U.S. partners. With the exception of Taiwan’s foundries, most of its firms are likely to continue to experience extreme pricing pressure, forcing them to respond by offshoring much of their lower-end work to China. INFRASTRUCTURE To ENABLE MULTIFACETED INNOVATION Information technology plays a pervasive and indispensable role in the United States. As IT becomes almost ubiquitous, Americans use it in increasingly sophisticated ways for work, family life, and entertainment. For example: • By 2006, almost 70 percent of adult Americans (18 and older) owned a desktop computer, and 30 percent had a laptop. • In 2001, the Apple iPod was introduced, forever changing the music and entertainment landscape: by 2006, 20 percent of adult Americans had an iPod. • Over the past decade, cellular telephones and handheld devices have launched a new form of “mobile” communications, connecting people through voice and data applications no matter where they are: in 2006, 73 percent of adult Americans had cell phones, and 11 percent had a handheld device.127 126Clair Brown and Greg Linden, “Semiconductor Engineers in a Global Economy,” in Na- tional Academy of Engineering, The Offshoring of Engineering: Facts, Unknowns, and Potential Implications, The National Academies Press, Washington, D.C., 2008, pp. 149-178. 127Data on the percentage of Americans reporting that they had specific technology are

The changing landscape: 1995-2007 89 These networking and device technologies are used for more than entertainment and desktop or laptop computing, however. Applications areas such as transportation, banking and financial services, and health care have been greatly impacted by use of IT. Information technology is now found in dishwashers, cars, lasers, medical equipment, smartcards, and numerous other devices and machines. Critical broadband connec- tivity connects these endpoints, keeping them working together to create business and consumer value. The network itself continues to evolve, just as do the devices and software that it connects—the infrastructure gets faster, cheaper, and more reliable, and devices are becoming smaller and multimodal.128 In the eyes of most Americans, these technologies are not only indis- pensable to the nation’s business operations and their derived productiv- ity, but also facilitate the ongoing learning and creativity of U.S. citizens. Multifaceted Innovation Information technology innovation no longer happens only in uni- versity or corporate laboratories. Customer-created value is increasingly prominent: IT consumers are leveraging research, innovating, and creat- ing value by combining networking hardware, software, and devices into novel solutions and businesses (see Figure 3.2). In 1995, supplier-created value through technological product innovations in information technol- ogy predominated. However, this pattern has been changing, as custom- ers are increasingly creating value through IT application innovations in industries including health care, professional services, financial services, manufacturing, retail, media and publishing, and education.129 As a result of the co-evolution of business and IT, the IT R&D ecosystem is becoming increasingly linked with R&D in the wider global economy. The significance of multifaceted innovation to IT has increased sig- nificantly during the study period 1995 to 2007. During this period, IT has permeated almost all aspects of business and society. One consequence of this pervasiveness is that the market-facing, customer-involved aspects of IT are growing very fast. This means, among other things, that “IT jobs” are changing—now often requiring customer-specific, market-specific, and business-specific expertise, not just technology-specific expertise. IT from Pew Internet and American Life Project Survey, April 2006, available at http://www. pewinternet.org/pdfs/PIP_ICT_Typology.pdf; accessed October 18, 2007. Pew surveyed 4,001 Americans 18 years of age and older by telephone. 128Information Technology and Innovation Foundation, Digital Prosperity: Understanding the Economic Benefits of the Information Technology Revolution, Washington, D.C., March 13, 2007. 129David Moschella, “Aligning R&D with Industry Change,” presentation to the commit- tee, Boston, Mass., April 19, 2007.

90 assessing the impacts of changes in the it R&D ecosystem 100 Customer-Created Value Percentage of Value Supplier-Created Value 0 1995 2010 FIGURE 3.2  Shift from supplier-created to customer-created value, 1995-2010. SOURCE:  Adapted from David Moschella, “Aligning R&D with Industry Change,” presentation to the committee, Boston, Mass., April 19, 2007. Figure 3-2.eps is enabling new products and services, and innovation in IT also includes innovations in these products and services. The National Research Coun- cil expects to embark on a congressionally mandated project looking at education, training, and research dimensions of innovation in IT-enabled services.130 Indeed, revenues in the IT sectors are increasingly coming from ser- vices (including software maintenance) rather than from product sales. According to Michael Cusumano, software product companies and types of firms experiencing the shift toward services face three kinds of chal- lenges: identifying the best revenue mix (of products and services, bearing in mind that products drive service revenues), creating service offerings that can make a firm’s products less commodity-like, and making service 130 Section 1005 of Public Law 110-69 (the America COMPETES Act of 2007) calls for the Office of Science and Technology Policy, through the National Academies, to conduct a study and to report to Congress on “how the Federal Government should support, through research, education, and training, the emerging management and learning discipline known as service science,” which is defined as “curricula, training, and research programs that are designed to teach individuals to apply scientific, engineering, and management disciplines that integrate elements of computer science, operations research, industrial engineering, business strategy, management sciences, and social and legal sciences, in order to encourage innovation in how organizations create value for customers and shareholders that could not be achieved through such disciplines working in isolation.”

The changing landscape: 1995-2007 91 delivery more efficient (through reuse of software components, standard- ized process frameworks and training, and automation of services). Thus both suppliers and their customers have ample new opportunities to inno- vate in underlying technologies, in product- and service-delivery models, in new business models, and in new product and service offerings. 131 U.S. leadership in many IT-related markets is under competitive pres- sure. The changing locus of IT innovation, including customers as well as university or corporate laboratories, makes demand leadership by U.S. consumers—that is, having consumers that are among the most techno- logically sophisticated in the world, with leading-edge product require- ments—increasingly important to the U.S. global competitiveness in IT. With multifaceted IT innovation, the rationale is that the most dynamic IT companies will ultimately be in countries with the most demanding IT customers.132 Thus if a nation’s users are not global lead users—requiring and using the most advanced IT functionalities—then in the market seg- ments where their demand lags, that nation’s user-driven IT innovation also will lag.133 U.S. consumers are not at the leading edge in important market seg- ments such as mobile telephones and wireless services.134 In part, this has to do with the fact that the wireless infrastructure in this country trails, in coverage and speed, the infrastructure in the European Union, Japan, and now increasingly, China.135 It is also increasingly evident that U.S. consumers are not the leading adopters of new wireless services such as 131As Michael A. Cusumano discusses, the shift to services on the part of traditional IT-product companies creates new opportunities for innovation, but it also creates new challenges for dedicated IT-services companies. See Michael A. Cusumano, “The Changing Software Business: Moving from Products to Services,” IEEE Computer, January 2008, pp. 20-27. 132For a recent survey of the literature on the adoption and diffusion of IT in businesses, in- cluding “co-invention,” see Chris Forman and Avi Goldfarb, “Diffusion of Information and Communication Technologies to Businesses,” in Terry Hendershott, ed., Handbook of Econom- ics and Information Systems, Elsevier, 2006, working paper version available at http://papers. ssrn.com/sol3/papers.cfm?abstract_id=896750#PaperDownload; accessed March 24, 2008. 133For development of the lead-user concept, see, for example, Eric von Hippel, “Lead U ­ sers: A Source of Novel Product Concepts,” Management Science 32(7):791-805, 1986; and Glen L. Urban and Eric von Hippel, “Lead User Analyses for the Development of New Industrial Products,” Management Science 34(5):569-582, 1988. 134Recent developments in the U.S. wireless market are signs of significant innovation, including the 3G iPhone from Apple, new competitor products by such vendors as Samsung and LG, and Google’s entry into the mobile phone arena with its Android platform. 135According to the European Union’s (EU’s) 12th report on telecommunications mar- kets, mobile penetration stood at 103 percent, and the EU had overtaken Japan with the largest population of 3G subscribers, with 45 million as of the end of 2006; available at http://ec.europa.eu/information_society/newsroom/cf/itemlongdetail.cfm?item_id=3304; accessed August 28, 2007.

92 assessing the impacts of changes in the it R&D ecosystem ring tones, ringback tones, games, and payment services. The direction of causality is unclear, but infrastructure clearly plays a fundamental role. Both consumer demand and enabling infrastructure are necessary: for example, it is pointless to demand video on a cell phone if the wireless infrastructure cannot support it, or to expect home health care delivery offerings to prosper if broadband penetration is low (see the subsection below entitled “Broadband Speeds and Capabilities” for examples of bandwidths needed to support particular functionalities). Network Infrastructure and Innovation Leadership An environment of leading-edge users of technology creates the essential context for technology’s next wave and its effective applica- tion. In such an environment, all sectors of society, including consumers, businesses, and governments, exploit and make best use of advanced information technology. However, as more leading-edge deployments of IT rely on mobility and a multimedia-capable infrastructure, it appears that the key IT suppliers in these markets will tend to focus their efforts on populations outside the United States because these markets are growing and because the infrastructures are better able to support users of these leading-edge technologies. This focus will, in turn, help those users grow in sophistication and comfort with the technology, surpassing users in the U.S. domestic markets. A situation in which U.S. consumers and users become increasingly less demanding in terms of product features and capabilities is cause for concern, because IT innovation is, increasingly, occurring at the “edge”—through user-driven application innovation. United States Behind in IT Deployment in Some Domains While the United States has long led the world as the largest IT mar- ket, thereby commanding the attention of the leading global providers of IT products and services, other countries are increasingly taking the lead in deploying IT for certain domains ahead of the United States. Accord- ing to Morgan Stanley Research, “The large U.S. telecom operators are well behind their European peers in regard to fixed/mobile convergence, due to both structural and system issues.”136 In addition to mobile carrier infrastructure, other areas in which the U.S. failure to deploy IT intel- ligently is causing our society to fall behind in important ways include broadband infrastructure deployment, health care, and homeland secu- 136Mark Shuper, Adnaan Ahmad, Simon Flannery, Nick Delfas, Scott Coleman, Vance Edelson, and Franklin Fu, Telecommunications 4G: Still the Early Days (WiFi/WiMax in Focus), Morgan Stanley Research Telecommunications Report, August 2006.

The changing landscape: 1995-2007 93 rity. For example, the Markle Foundation’s 2002 report Protecting America’s Freedom in the Information Age describes the importance of first-rate infor- mation collection, analysis, communications, and sharing for purposes of countering threats from terrorism and weapons of mass destruction, as well as pointing out how the U.S. failure to mobilize and deploy IT resources effectively harms response capabilities.137 In the past, the U.S. government has played a strong role in estab- lishing U.S. IT demand leadership. From the 1960s through the 1980s, the U.S. government played a fundamental role in the development of numerous fields of information technology—both as a sponsor of broad- based research and as a lead customer in emerging markets.138 During this period, often through military and space programs, the U.S. govern- ment served as a demand leader—a first customer for new commercial products that promised orders large enough to sustain investment in new products and processes.139 Government has also served as a builder of infrastructure in advance of wider demand, notably for networking in research and education. However, as commercial markets have outpaced federal procure- ments, the federal government’s role in shaping and sustaining the IT R&D ecosystem has diminished. For example: • As IT industries have matured and commercial demand for IT has soared, the government has ceased to be the “lead” customer (in terms of 137See Markle Foundation, Protecting America’s Freedom in the Information Age, New York, N.Y., October 2002, available at http://www.markle.org/downloadable_assets/nstf_full. pdf; see also Markle Foundation, Creating a Trusted Information Network for Homeland Security, New York, N.Y., December 2003, at http://www.markle.org/downloadable_assets/nstf_ report2_full_report.pdf; both accessed August 28, 2007. See also Jonathan Marino, “DHS Tech Chief Wants Broadband for First Responders,” Government Executive, March 15, 2007, avail- able at http://www.govexec.com/dailyfed/0307/031507j2.htm; accessed October 18, 2007. 138For example, research supported by the Defense Advanced Research Projects Agency’s (DARPA’s) Information Processing Techniques Office (IPTO) from 1962 through the mid- 1980s led to developments including time-sharing, interactive computer graphics, net- working, integrated circuit design, and intelligent systems. For a comprehensive history of DARPA’s IPTO and its style of “managing for innovation,” see Arthur L. Norberg and Judy E. O’Neill, with contributions by Kerry J. Freedman, Transforming Computer Technology: Information Processing for the Pentagon, 1962-1986, Johns Hopkins University Press, Baltimore, Md., 1996. See also National Research Council, Funding a Revolution: Government Support for Computing Research, National Academy Press, Washington, D.C., 1999. 139Federal procurement of integrated circuits for NASA’s Apollo spacecraft and for the Minuteman II missile guidance system sparked and sustained early industry investments in manufacturing capacity and encouraged commercial markets for integrated circuits. A notable example of a civilian agency as lead customer is the Census Bureau’s purchase of the first Univac computer in 1951. See also National Research Council, Funding a Revolution: Gov- ernment Support for Computing Research, National Academy Press, Washington, D.C., 1999.

94 assessing the impacts of changes in the it R&D ecosystem cutting-edge needs or dominant purchasing power) for general-purpose hardware or software. The military continues to be a lead customer in certain specific areas (for example, extremely large scale, cyberphysical weapons systems), and in these it faces a struggle to ensure adequate internal and contractor capabilities.140 • In some cases, where the government’s requirements had been perceived as beyond then-current commercial offerings, the government chose to “make” rather than “buy” needed technologies. Sometimes, the government finds it difficult to—or does not—appreciate when the deci- sion crossover between “make” and “buy” has occurred. As a result, some agencies must now struggle to maintain decades-old and obsolete, but mission-critical, technologies while also attempting to modernize these systems using state-of-the-practice commercial technologies. 141 • Although the U.S. government was rated a leader in readiness for e-government in 2003,142 this may not translate into the same kind of lead role that the Defense Advanced Research Projects Agency (DARPA) played in prior decades.143 • The U.S. federal government generally does not stand at the fore- front in terms of innovative IT use, and federal spending on IT does not dominate the commercial marketplace. Thus the government does not serve as an effective “lead customer” to spur development of new and innovative commercial technologies and products. • In cooperation with the business school INSEAD, the World Eco- nomic Forum produces The Global Information Technology Report, which 140This is the subject of an ongoing National Research Council (NRC) study on advancing software-intensive system producibility and of an NRC workshop report: National Research Council, Software-Intensive Systems and Uncertainty at Scale, The National Academies Press, Washington, D.C., 2007. For a discussion of the implications of increased Department of Defense reliance on commercial software during a period of increasing globalization in IT industries, see also Defense Science Board, Report of the Defense Science Board Task Force on Mission Impact of Foreign Influence on DoD Software, Office of the Undersecretary of Defense for Acquisition, Technology, and Logistics, Washington, D.C., September 2007. 141See National Research Council, The Social Security Administration’s E-Government Strategy and Planning for the Future, The National Academies Press, Washington, D.C., 2007. 142In 2003, a United Nations survey found that the United States led the world in “e- government readiness,” followed by Sweden, Australia, Denmark, and the United Kingdom. Readiness was measured by a composite index based on an assessment of Web sites, tele- communications infrastructure, and human resource endowment (including literacy). See United Nations, UN Global E-Government Readiness Report: UN Global E-Government Survey 2003, available at http://www.unpan.org/egovernment3.asp; accessed July 16, 2007. 143“DARPA funding of advanced technologies, particularly in Information Technology (IT), has had enormous impact, although largely on platform technologies that had wide and profound spillovers.” See National Research Council, Innovation Policies for the 21st Century: Report of a Symposium, The National Academies Press, Washington, D.C., 2007, footnote 2, p. xiv.

The changing landscape: 1995-2007 95 ranks 122 countries on the basis of their “networked readiness.” This is a metric that the Forum uses to measure the countries’ preparation to participate in and benefit from developments in information technology. In the 2006-2007 report, the networked readiness ranking for the United States was seventh place; the United States had been in first place in the 2005-2006 rankings. The drop in the U.S. ranking was attributed to “rela- tive deterioration of the political and regulatory environment.”144 In its 2007-2008 report, however, the World Economic Forum raised the United States’ networked readiness ranking to fourth overall, after Denmark, Sweden, and Switzerland. The new report placed particular focus on the role of networked readiness in spurring innovation. The reported U.S. strengths included availability of capital and the quality of U.S. R&D insti- tutions; the reported weaknesses included cost and speed of broadband connectivity.145 Comparing Aspects of Broadband in the United States and Abroad Compared with the more highly regulated environment of past de­cades, the current telecommunications market environment in the United States has yielded many consumer benefits. However, these ben- efits have not accrued evenly. By the early years of the 21st century, although broadband was regarded as a national and local imperative, there was substantial geographical variation in the nature of broadband competition, broadband was not available everywhere, and investments 144The World Economic Forum’s national Networked Readiness Indicator (NRI) has three components: the environment for IT offered by the country; the readiness of the country’s individuals, businesses, and governments; and the usage of IT among these stakeholders. The 10 top-ranked countries were Denmark, Sweden, Singapore, Finland, Switzerland, the Netherlands, the United States, Iceland, the United Kingdom, and Norway. These countries all had NRI scores between 5.71 and 5.42. By comparison, France ranked 23rd, with a score of 4.99, and Mexico ranked 49th, with a score of 3.91. (However, the United States was cited for maintaining its “primacy in innovation, driven by one of the world’s best tertiary education systems and its high degree of cooperation with the industry as well as by the extremely efficient market environment.”) See World Economic Forum, “Denmark Climbs to the Top in the Rankings of the World Economic Forum’s Global Information Technology Report 2006-2007,” Press Release, available at http://www.weforum.org/en/media/Latest%20 Press%20Releases/gitr_2007_press_release; accessed July 18, 2007. 145World Economic Forum, The Global Information Technology Report 2007-2008, available at http://www.weforum.org/en/initiatives/gcp/Global%20Information%20Technology%20 Report/index.htm; accessed April 9, 2008. Some observers reportedly were skeptical of the improvement in the U.S. ranking owing to their concerns about U.S. broadband capabilities, penetration, adoption, and costs. See John Markoff, “Study Gives High Marks to U.S. In- ternet,” New York Times, April 9, 2008, available at http://www.nytimes.com/2008/04/09/ technology/09internet.html?ex=1208404800&en=5625fba016b5acbf&ei=5070&emc=eta1; a ­ ccessed April 9, 2008.

96 assessing the impacts of changes in the it R&D ecosystem in additional facilities and performance improvements were uncertain.146 In that environment, although the United States was a world leader in computer usage, it was already lagging in broadband connectivity (espe- cially in the areas of speed and price—see Table 3.2) compared with other countries (albeit those with geographic, population-density, and indus- trial policy characteristics different from those of the United States). A number of international rankings show that the United States lags in international comparisons. For example, IDC’s “Information Society Index” (ISI) measures the ability of 53 nations to participate in the infor- mation revolution. To construct the ISI, IDC includes 15 variables grouped into four types of infrastructure indexes: social, Internet, computer, and telecommunications (telecom) infrastructures. In 2003, the United States ranked first in IDC’s computer index, but only 20th in the telecom index, which included the number of broadband households.147 Another inter- national ranking, by the International Telecommunications Union, based on broadband subscribers per 100 people, put the United States in 20th place in 2006, after a steady decline from 3rd place in 1999.148 Table 3.2 presents a snapshot of the United States’ uneven interna- tional standing in broadband in 2007: according to OECD data, it leads in total number of subscribers, is in the middle of the 10 countries listed in terms of per capita penetration, and is far behind in advertised down- load speed (at relatively high prices). However, like some of their foreign counterparts, U.S. carriers have continued to deploy combination service offerings and pricing arrangements (for example, bundling television, telephone, and data services in one cable or fiber-optic phone offering), and therefore prices and capabilities are likely to continue to improve in at least some areas of the United States. The goal of more-ubiquitous, lower-cost, and higher-speed broad- band deployment149 has been the focus of significant analysis and advo- 146National Research Council, Broadband: Bringing Home the Bits, National Academy Press, Washington, D.C., 2002; discussion of findings on pp. 13, 18, and 21. 147In 2003, the top 10 countries in IDC’s composite ISI rankings were Denmark, Sweden, United States, Switzerland, Canada, Netherlands, Finland, Korea, Norway, and the United Kingdom. The IDC’s computer index includes PCs per household, IT spending as a fraction of GDP, IT services’ contribution to GDP, and software spending; the telecom index includes the number of broadband households, wireless subscribers, and handset shipments. See IDC, “IDC’s Information Society Index,” available at http://www.idc.com/groups/isi/ main.html; accessed July 18, 2007. 148J. Windhausen, Jr., A Blueprint for Big Broadband, EDUCAUSE White Paper, January 2008, p. 12, citing International Telecommunications Union data, available at http://www. educause.edu/ir/library/pdf/EPO0801.pdf; accessed March 13, 2008. 149For technical, regulatory, and policy analyses of broadband, see National Research Council, Broadband: Bringing Home the Bits, National Academy Press, Washington, D.C., 2002.

The changing landscape: 1995-2007 97 Table 3.2 A Snapshot Comparison of Broadband in 10 Countries in 2007 Average Advertised Total Number of Broadband Average Number of Broadband Download Monthly Broadband Subscribers Speed Cost of Subscribers per 100 (megabits Broadband Country (million) Inhabitants per second) (U.S. $) United States 66.2 22.1 8.9 53 Japan 27.2 21.3 93.7 34 Germany 17.5 21.2 9.2 NA Korea 14.4 29.9 43.3 42 United Kingdom 14.4 23.7 10.6 33 France 14.3 22.5 44.2 37 Italy   9.3 15.8 13.1 NA Canada   8.1 25.0 7.8 51 Spain   7.5 17.0 6.901 NA Netherlands   5.5 33.5 5.312 39 NOTE: NA, not available. Source: Based on data of the Organisation for Economic Co-operation and Development presented in J. Windhausen, Jr., A Blueprint for Big Broadband, EDUCAUSE White Paper, January 2008, pp. 20-21; 23-24, available at http://www.educause.edu/ir/library/pdf/ EPO0801.pdf. cacy. In January 2002, for example, TechNet, a group of Silicon Valley chief executive officers, proposed that the President and policy makers “make broadband a national priority and set a goal of making an affordable 100-megabits per second broadband connection available to 100 million American homes and small businesses by 2010.”150 A June 2007 report from the Information Technology and Innovation Foundation (ITIF) uses an externalities argument to make its case that government action is needed to advance broadband deployment, because market forces will not be sufficient: First, it [broadband] is a not just a consumer technology like the iPod or Blu-Ray player, it is “prosumer” technology that is enabling consumers to also be producers who contribute to economic growth and innova- tion. Second, it exhibits positive network externalities where the benefits from broadband adoption accrue not just to individual consumers, but to other broadband users and society as a whole. Because of this the 150See TechNet, A National Imperative: Universal Availability of Broadband by 2010, January 15, 2002, available at http://www.technet.org/resources.dyn/2002-01-15.64.pdf; accessed June 27, 2007.

98 assessing the impacts of changes in the it R&D ecosystem social returns from investing in more broadband exceed the private returns of companies and consumers. As a result, market forces alone will not generate the societally optimal level of broadband, at least for the foreseeable future. In markets like this, public policies—in this case a proactive national broadband strategy—are needed to maximize overall societal welfare.151 Additionally, based on 2006 OECD data, the ITIF found that the United States had fallen to rank 12th behind countries including Korea, Japan, and Iceland. A 2007 report of the National Telecommunications and Information Administration (NTIA), Broadband in America, describes federal efforts toward the vision of “universal, affordable access” to broadband technol- ogy: these include Federal Communications Commission (FCC) efforts to modify regulations in order to provide incentives for network invest- ments by local telephone companies and to stimulate facilities-based investments by other providers, support for cable franchise reforms, and more timely and cost-effective access to rights-of-way on federal land. 152 Using data from various sources, the NTIA reported large increases in various types of high-speed network access (via telephone lines and cable, as well as high-speed wireless) and decreases in prices, over the period from 2001 to 2007. Significantly, however, the NTIA report notes that “the lack of a sin- gle authoritative data set makes it difficult to establish with certainty whether broadband penetration has become ubiquitous, and this Report acknowledges the benefits of better data gathering tools.”153 In part, the piecemeal nature of the U.S. data compared with the data available for some other countries naturally reflects the multiplicity of federal, state, and local policies and regulatory regimes for different types of technolo- gies and providers, as well as the large and growing number of providers (see Table 3.3). Nevertheless, data limitations make it difficult to piece together a complete, current snapshot of broadband in the United States or to evaluate the various claims regarding progress—or lags—in broad- band availability. Moreover, the often wide differences between available “broadband” speeds in the United States and foreign counties complicate direct comparisons. 151Robert D. Atkinson, The Case for a National Broadband Policy, The Information Technology and Innovation Foundation, Washington, D.C., June 2007. 152National Telecommunications and Information Administration, Networked Nation: Broadband in America, U.S. Department of Commerce, Washington, D.C., January 2008, pp. i, ii, available at http://www.ntia.doc.gov/reports/2008/NetworkedNationBroadbandin America2007.pdf; accessed March 13, 2008. 153Ibid., p. 12.

The changing landscape: 1995-2007 99 Table 3.3 Number of Providers of High-Speed Lines Nationwide in the United States, 1999-2006, by Technology (over 200 kilobits per second in at least one direction) Number of Providers Cable All Month, Year ADSLa Modem Otherb Totalc December 1999   28   43   65 105 December 2000   68   39   87 136 December 2001 117   59 122 203 December 2002 178   87 169 299 December 2003 274 110 246 432 December 2004 352 147 312 552 December 2005 820 242 835 1,347 December 2006 862 278 882 1,397 NOTES: According to the National Telecommunications and Information Administration, data through December 2004 include only providers with at least 250 lines per state, which were the only ones required to file; some historical data have been revised.   According to the 2002 report of the National Research Council entitled Broadband: Bring- ing Home the Bits (National Academy Press, Washington, D.C., 2002, p. 63), 200 kilobits per second is not adequate to support a single, TV-quality video stream to each house.   aADSL, or asynchronous digital subscriber line, is carried over copper telephone lines. Because it provides essential infrastructure, broadband constitutes a foundation for leader- ship elsewhere. Attention here could produce benefits in a number of other areas, including health care (for example, access to broadband facilitates the transfer and analysis of elec- tronic patient records and test results, particularly imaging). Note that the Federal Com- munications Commission has started a pilot funding program for a nationwide, broadband network dedicated to health care. See “Rural Health Care Pilot Program,” available at http://www.fcc.gov/cgb/rural/rhcp.html; accessed October 18, 2007.   b“All Other” includes synchronous digital subscriber line (SDSL), traditional wireline, fiber, satellite, fixed and mobile wireless, and power line.   cTotal is not simply the sum of the first three columns because some providers offer services using multiple technologies. Source: Data from National Telecommunications and Information Administration, Net- worked Nation: Broadband in America, U.S. Department of Commerce, Washington, D.C., 2008, Table 1, based on data from Federal Communications Commission, High-Speed Services for Internet Access: Status as of December 31, 2006, Washington, D.C., October 2007, Table 7. Unlike the United States, Korea and Japan are small in area, with political institutions that favor a government role in industrial policy. While overall comparisons among countries are difficult, relative rank- ings in broadband penetration, speeds, and costs are nonetheless relevant because of the linkages between enabling infrastructure, demand leader- ship, and innovation leadership. For example, although the household penetration (fraction of households that subscribe to a broadband service) of broadband in Korea in 2007 was 90 percent, in the United States it was

100 assessing the impacts of changes in the it R&D ecosystem only 51 percent. The United States also lags in penetration per 100 inhabi­ tants (see Table 3.2). In 2007 the average bandwidth was over 40 Mbps in Korea; it was under 10 Mbps in the United States; the average cost per 1 megabit of capacity was under $1 per month in Korea; it was almost $6 in the United States (see Table 3.2). Korea’s high-speed infrastructure is widely credited with enabling its inhabitants to attain demand leadership in content-rich online games. Japan is pursuing a very aggressive strategy of broadband deploy- ment. It reportedly had the world’s fastest broadband service in 2007 (see Table 3.2), a speed (on average, 93.7 Mbps) that enables Japanese con­ sumers to watch full-screen, broadcast-quality television over the Inter- net. Japan’s broadband lead over the United States is attributed in part to ­ better physical infrastructure (newer and better telephone wires and shorter distances between the central office and homes); DSL in Japan is often 5 to 10 times as fast as the services widely offered by U.S. cable pro- viders. However, Japanese industrial policy also plays a role: the Japanese government used subsidies, tax incentives, and regulation to promote high-speed broadband deployment: • Government subsidies and tax incentives reportedly spurred Nip- pon Telegraph and Telephone Corp.’s (NTT’s) nationwide build-out of fiber-optic lines (offering connection speeds of up to 100 megabits per sec- ond) to about 8.8 million Japanese homes. NTT, Japan’s largest telephone company, was once government-controlled. • Government regulation required large telephone companies (NTT, for example) to open up their copper wire networks to small Internet pro- viders at prices that allowed these new broadband companies to charge as little as $22 a month for a DSL connection faster than almost all U.S. broadband services. These levels of broadband service are enabling the development of a number of valuable new applications, such as low-cost, high-definition teleconferencing for telemedicine and advanced telecommuting. 154 A fundamental step to being the world leader in information tech- nology use is for the United States to deploy world-class broadband connectivity aggressively over the next decade. The United States cur- rently lags behind other nations such as Japan and Korea in upgrading and deploying national broadband connectivity. Setting, and reaching, a highly ambitious target—such as making 1,000 megabits per second 154See Blaine Harden, “Japan’s Warp-Speed Ride to Internet Future,” Washington Post, August 29, 2007, p. A01, available at http://www.washingtonpost.com/wp-dyn/content/ article/2007/08/28/AR2007082801990_pf.html; accessed August 29, 2007.

The changing landscape: 1995-2007 101 broadband connectivity available to 100 million American homes and small businesses by 2020—would enable the United States to leap well ahead of other countries in this area and to hold that lead. 155 Governments can use economic incentives and targeted regulations to promote higher-speed connectivity across a common physical infra- structure. By using multiple wavelengths or colors, a single fiber today is able to carry 1 to 10 terabits of data.156 U.S. terrestrial fiber networks have large amounts of “dark” (unused) fiber, and many fibers already lit could accommodate additional colors.157 However, a large obstacle remains: the deployment of fiber or the installation of other upgrades to the “last mile” to connect all the endpoints (homes, businesses, government agen- cies, and other organizations) to the national networks. In the United States, the connectivity landscape is in part a product of historical policy goals (such as universal access for telephony) and the structure of U.S. economic regulation. There is merit in considering models for broadband d ­ eployment—for example, models of companies competing in the value- added services market using a common physical infrastructure,158 or models whereby facilities-based competition is fostered.159 Value-added services that can benefit from gigabit connectivity include movies on demand, multimedia Web browsing, many-to-many video communica- tions, news groups, and so forth. However, the question remains as to who makes the infrastructure investments and who extracts the value of these services. In the United States, the complex system of federal, state, and local governance and regulations can present numerous transactional bottle- necks, such as right-of-way restrictions and content franchising (for exam- ple, for video), to pursuing such approaches. These may tend to favor the 155This target is more ambitious than TechNet’s proposal for accelerating broadband deployment and demand, which called for 100 megabit-per-second connectivity by 2010. See “Accelerating Broadband Deployment and Demand,” available at http://www.technet. org/issues/broadband/; accessed September 7, 2007. A goal of gigabit connectivity would be useful in helping the United States leapfrog Japan and other nations now moving ahead in broadband deployment. 156See “Introducing DWDM [Dense Wavelength Division Multiplexing],” http://www. cisco.com/univercd/cc/td/doc/product/mels/cm1500/dwdm/dwdm_fns.htm; accessed September 7, 2007. 157TeleGeography Research, “Global Bandwidth Research Service: Executive Summary,” Washington, D.C., 2008, available at http://www.telegeography.com/products/gb/index. php; accessed October 31, 2008. 158An inexact analogy would be the federal government paying for an interstate highway system and the private sector creating products (such as cars, gas stations, and motels) that benefit from the use of this infrastructure. 159See National Research Council, Broadband: Bringing Home the Bits, National Academy Press, Washington, D.C., 2002.

102 assessing the impacts of changes in the it R&D ecosystem incumbents and slow overall progress toward attaining higher-speed, lower-cost broadband deployment that can support data-rich IT applica- tions and services and enable leading-edge, ­customer-driven innovation by U.S. consumers. Broadband Speeds and Capabilities With respect to broadband, how fast is fast enough? That is, what bandwidth target is desired in order to enable multifaceted innovation? The “answer” is actually a moving target. Consequently, in the 2002 National Research Council report Broadband: Bringing Home the Bits, the Committee on Broadband Last Mile Technology did not set bandwidth- specific definitions for what constituted “broadbands,” and it deliber- ately did not set specific bandwidth targets for policy makers. Instead, that committee established a functional definition: “Broadband services should provide sufficient performance—and wide enough penetration of services reaching that performance level—to encourage the development of new applications.”160 Furthermore, that committee recommended a more coherent, consistent broadband policy framework that is service- oriented, rather than being technology-centric.161 An important consideration in thinking about broadband leadership and the question of what bandwidth to “target” is the fact that broadband data rates considered adequate a few years ago are no longer sufficient to support new applications and services.162 Higher-speed services attract more customers because they are more useful for high-data-rate applica- tions (such as video). A higher-speed infrastructure stimulates multifac- eted innovation. In January 2008, the California Broadband Task Force (CBTF) pub- lished its final report, The State of Connectivity: Building Innovation Through Broadband.163 The CBTF recommendations included building out “high speed” broadband infrastructure for all Californians, as well as promoting innovative uses of broadband technology. The CBTF adopted a working definition of broadband that includes a basic minimum speed (expected to increase over time) of 512 kbps.164 Table 3.4, adapted with minor stylistic 160Ibid., p. 80. 161Ibid., pp. 32-33. 162Ibid.; see, especially, Ch. 2 for a discussion of then-current broadband technologies, speeds, and capabilities (as well as economic, regulatory, and policy factors). 163California Broadband Task Force, The State of Connectivity: Building Innovation Through Broadband, January 2008, available at http://www.calink.ca.gov/taskforcereport/; accessed March 17, 2008. 164Ibid., pp. 8, 12..

The changing landscape: 1995-2007 103 Table 3.4 Bandwidth Ranges Corresponding to Advanced Applications and Services Bandwidth Range Applications and Services Enabled 500 kbps–1 Mbps Voice over IP [Internet Protocol] SMS [short message service] Basic e-mail Web Browsing (simple sites) Streaming Music (caching) Low-Quality Video (highly compressed) 1 Mbps–5 Mbps Web Browsing (complex sites) E-mail (larger-size attachments) Remote Surveillance IPTV-SD (1-3 channels) [standard definition Internet Protocol   television] File Sharing (small/medium) Telecommuting (ordinary) Digital Broadcast Video (1 channel) Streaming Music 5 Mbps–10 Mbps Telecommuting (converged services) File Sharing (large) IPTV-SD (multiple channels) Switched Digital Video Video on Demand SD Broadcast SD Video Video Streaming (2-3 channels) HD [High-Definition] Video Downloading Low-Definition Telepresence Gaming Medical File Sharing (basic) Remote Diagnosis (basic) Remote Education Building Control and Management 10 Mbps–100 Mbps Telemedicine Educational Services Broadcast Video SD and Some HD IPTV-HD [high-definition Internet Protocol television] Gaming (complex) Telecommuting (high-quality video) High-Quality Telepresence HD Surveillance Smart/Intelligent Building Control continued

104 assessing the impacts of changes in the it R&D ecosystem Table 3.4 continued Bandwidth Range Applications and Services Enabled 100 Mbps–1 Gbps HD Telemedicine Multiple Educational Services Broadcast Video Full HD Full IPTV Channel Support Video on Demand HD Gaming (immersion) Remote Server Services for Telecommuting 1 Gbps–10 Gbps Research Applications Telepresence Using Uncompressed High-Definition Video   Streams Live Event Digital Cinema Streaming Telemedicine Remote Control of Scientific/Medical Instruments Interactive Remote Visualization and Virtual Reality Movement of Terabyte Data Sets Remote Supercomputing Source: Adapted from table entitled “What Is Broadband,” p. 12, California Broadband Task Force, The State of Connectivity: Building Innovation Through Broadband, January 2008, available at http://www.calink.ca.gov/taskforcereport/; accessed March 17, 2008. changes from the CBTF report, illustrates the types of applications and services made feasible by increasing bandwidth. SUMMARy In this chapter, the intention of the committee has been to illuminate the complex story of the evolution of the U.S. IT R&D ecosystem during the 1995-2007 period. First, it summarized the tumultuous business and technological changes experienced in the IT industry since 1995. The IT R&D ecosystem was affected by business transformations as the Inter- net was commercialized. In the process, the world experienced the larg- est venture capital investment bubble in history and an accompanying dramatic stock market bubble. The bubble may not have been entirely negative, because major new firms were created and the ways that people work and play were transformed. However, the collapse of the bubble did lead to a massive reduction in venture capital investing that some believe significantly retarded the commercialization of information technologies. Also, the collapse of the bubble may have discouraged students from entering the computer science and computer engineering fields, possibly leading to longer-term labor shortages.

The changing landscape: 1995-2007 105 The committee then turned its attention to the evolution of major platforms (such as Web 2.0, open-source development, new mobile access devices, and services executing within Internet data centers) and to the evolution of the major component sectors of semiconductors, computers, and software. In technological terms, there were two extremely powerful major developments during the period of study: The first of these was the mass popularization of the Internet for purposes of business, uses as tools, and recreational use. The second was the rise of mobile telephony. Information technologies in this time period became ubiquitous. In purely technical terms, IT has permeated nearly every part of daily existence and knitted the world closer together. With this change came a globalization in which, for the first time in history, engineers even in developing nations became more capable of being integrated in the global economy. By dis- cussing India and China—two growing, potential IT industry giants—in particular, the committee places the situation of the U.S. IT R&D eco­ system into a global context. Today, it is no longer possible to understand the health and competitiveness of an isolated U.S. IT R&D ecosystem; it is now necessary to place it in a global context. Finally, the committee considered the multifaceted nature of IT inno- vation. IT innovation is no longer mainly supplier-driven. Increasingly, customers are creating value through application innovations. As these new applications and IT-enabled services grow in importance, IT workers will increasingly need more than just technology skills. They will need in-depth business- and market-related knowledge to leverage technology use and differentiate their products and services. For the United States to lead in this new environment, an appropriate network infrastructure is required: ubiquitous, higher-speed, and more-affordable broadband. With that as background for understanding the current state of the U.S. IT R&D ecosystem, the next chapter argues that the changes since 1995 have resulted in a globalized and fast-changing R&D ecosystem. If the United States does not navigate successfully in this global environ- ment, it will no longer enjoy a position at the center of technological change, one that it has enjoyed for the past decade or more.

Next: 4 A Globalized, Dynamic Information Technology R&D Ecosystem »
Assessing the Impacts of Changes in the Information Technology R&D Ecosystem: Retaining Leadership in an Increasingly Global Environment Get This Book
×
Buy Paperback | $62.00 Buy Ebook | $49.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The U.S. information technology (IT) research and development (R&D) ecosystem was the envy of the world in 1995. However, this position of leadership is not a birthright, and it is now under pressure. In recent years, the rapid globalization of markets, labor pools, and capital flows have encouraged many strong national competitors. During the same period, national policies have not sufficiently buttressed the ecosystem, or have generated side effects that have reduced its effectiveness. As a result, the U.S. position in IT leadership today has materially eroded compared with that of prior decades, and the nation risks ceding IT leadership to other nations within a generation.

Assessing the Impacts of Changes in the Information Technology R&D Ecosystem calls for a recommitment to providing the resources needed to fuel U.S. IT innovation, to removing important roadblocks that reduce the ecosystem's effectiveness in generating innovation and the fruits of innovation, and to becoming a lead innovator and user of IT. The book examines these issues and makes recommendations to strengthen the U.S. IT R&D ecosystem.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!