The Tide of Power - The Development and Prospects of Global Information Technology (1964-2010)

Author: (USA) David M. Gerstein, translated by Gao Jie et al.
Translator: Gao Xian, Gao Jie, Gao Duo
Country:
Publisher:
Publish Date: 2002-03-01
Features: As a major innovation in the development of personal computers, IBM's decision in 1981 to grant the key components of personal computers to Microsoft and Intel is now widely regarded as one of the most disruptive corporate decisions in business history. For over 60 years, IBM had meticulously built its monopoly through the personal computer industry. However, it was carelessly and suddenly relinquished to two companies that were only a tiny fraction of its size at the time. Clearly, IBM failed to recognize the strategic importance of personal computer and microprocessor technology. But everyone else in the information industry at the time was in the same situation. IBM's decision to use existing technology to launch quickly was almost universally praised, especially when IBM continued to dominate the personal computer industry in the early 1980s. After all, no major pioneer in personal computers—Apple, Commodore, or Tandy—manufactured its own microprocessor. In 1983, IBM's personal computers held an astonishing 70% share of the global market, with countless other suppliers vying for the remaining share. Although an IBM-compatible market was already emerging, it was not particularly alarming. William F. Zachmann, then the director of research at IDC, was one of the few who clearly and strongly believed that personal computers were not just another IBM plug-compatible market, such as hosts, disk and tape drives, printers, and terminals. If IBM did not control the key component technologies, its high-cost structure, slow time-to-market, and crippling internal bureaucratic infighting would quickly make it vulnerable to direct and open competition. Looking back now, it is clear that this view was generally ignored and often ridiculed before the late 1980s, but subsequent evidence has become undeniable. Even after the potential dangers of its decision became apparent, IBM and many others in the industry continued to believe that IBM would eventually reclaim what it had relinquished. Due to the significant and lasting impact of sellers such as Intel, Compaq, and especially Microsoft, almost all of these initiatives—PCXT/370, Topview user interface, Microchannel bus architecture, OS/2, WorkplaceOS, and newer PowerPC—have been disappointing or have proven completely unsuccessful. Given that most others in the industry viewed many of IBM's actions with the same enthusiasm as IBM itself, there is little need to blame IBM's specific leadership. Perhaps what must be said is that during the period when companies were constantly reevaluating their business models and questioning their fundamental market assumptions, IBM was severely punished for its inaction. IBM's decision to rely on Microsoft and Intel was precisely the kind of unconventional or out-of-the-box action that is often seen as necessary today. Bill Gates, for example, has carefully considered Microsoft's recommendations regarding various internet-based competitive approaches, whether solicited or unsolicited, such as Navigator from Netscape and Java from Sun, undoubtedly revealing the source of his own immense success.
Another source of Microsoft's power
IBM was not the only seller to inadvertently contribute to Microsoft's power. At times, it seemed as though the entire industry deliberately sought to push everything in Microsoft's direction. Consider the following three examples.
Apple's decision to make its Macintosh operating system non-proprietary allowed Microsoft to develop a similar graphical user interface over the years. By the time Apple recognized the value of its software assets, it was too late to change course. If it had acted earlier, the personal computer market might have been more evenly divided.
The potential merger between (Lotus) and Novell in the early 1990s could have become a formidable competitor to Microsoft. The merger failed primarily due to the arrogance of Lotus's Jim Manzi and Novell's Ray Noorda, who could not agree on who would lead the combined company. Novell continued to lose focus on its core networking business and made extremely unwise decisions, such as merging with WordPerfect. Eventually, Lotus was acquired by IBM.
Countless failed attempts over a decade to establish a Unix standard left Microsoft with ample time to develop Windows NT and position it as the standard for server operating systems. Only five years ago, many people found it laughable that Microsoft had control over the enterprise server environment. The only question now is how powerful that position will become. The Unix vendor group has largely abandoned the pursuit of full standardization. These events, as the primary source of Microsoft's compatibility processing power, have greatly benefited Microsoft and directly helped Intel.
This does not mean that Microsoft, Intel, and Novell have not done many outstanding things on their own. Microsoft has managed to extend the dominance of its operating system to applications, languages, tools, and other areas; owning the operating system alone does not guarantee this. Novell seized control of local network dominance without the help of IBM or anyone else. Intel is also unwilling to rest on its laurels. The company continues to invest heavily in development and manufacturing capabilities, keeping chip costs low while making it difficult for potential Intel-compatible sellers and other microprocessor architectures. In short, IBM gave Microsoft and Intel a huge opportunity to start, which they have fully exploited. Novell created its own opportunities but wasted much of its strength in the early 1990s, failing to recognize its true interests and trying to challenge Microsoft's dominant desktop position. Under new management leadership, the company has managed to return to its network software roots and shed its heavily discounted Unix and application businesses.
A consolidated software market has two main aspects of the software industry that attract far more attention than the other. First, some names must be clarified. Throughout the book, the term "software" primarily refers to specific computer programs—operating systems, utilities, applications, interfaces, protocols, and the like. It is generally not used to refer to what this book calls "content services"—music, video, text. Software consists of a set of instructions; content services revolve around some form of information, whether text, images, sound, video, or their combinations. Historically, from an analytical perspective, if software and content services can be evaluated separately, it is generally a good idea, as these two businesses often have fundamentally different market characteristics. The software market is often broadly horizontal in nature, while the content services market, though sometimes broad, is often fragmented and highly driven by individual consumer interests. The software market often has monopolistic oligopolies, whereas the content services market does not. However, on the World Wide Web and in the increasingly emerging consolidated industries, some of these distinctions will blur.
Still, in many fields, program instructions and information content will remain clearly separated, but increasingly, especially in interactive applications, program instructions and content services will become sufficiently intertwined to essentially merge into one. Microsoft and others often use the term "rich content" to describe this new form of information. This is a useful term to mark the new form of information. This new form means new competition. Who will develop rich content—the existing software vendors? The existing content service companies? Joint ventures between software and content service companies, such as between Microsoft and NBC? Or new rich content startups to replace existing operators? Consolidation will clearly lead to restructuring in many parts of the industry. The issue of rich content has attracted media attention, much like the hardware sector's discussions about PCs, network computers, and cable set-top boxes. Like hardware, when the software that manages the entire network infrastructure undergoes its own consolidation, equally significant changes may occur behind the scenes in the software industry.
Although the importance of computer software companies and major content service providers is easily understood, it is often forgotten how much software has driven the global telephone industry. Consider the current scope of computer software in building information—usernames, directory services, addresses, updates, security, metering, billing, and more. Our current telecommunications systems also have all the important equipment functions. For years, the telephone industry has invested heavily in future computer network management software, especially as integrated voice/data applications began to combine with payment, billing, and security systems. When user applications consolidate, competition will change again. While the computer industry has dominated discussions at internet engineering working groups so far, when telephone companies become the primary internet access providers, they will bring their technology and ambitions.
Can Intel's monopoly last? So far, the analysis of hardware has focused on the end-user device side of the industry, pointing out that U.S. leadership in this sector will soon face significant challenges. One might argue that this is not a major issue, as the core value-add of U.S. large hardware companies lies not in personal computer products (which anyone can assemble) but in fundamental semiconductor components. With Japanese and Korean companies dominating the global dynamic random access memory (DRAM) market, this analysis is primarily based on microprocessors. Most of the cost of personal computer hardware comes from microprocessors, DRAM, disk drives, and displays. While there are prominent markets for specialized semiconductors for graphics, video, audio, and other functions, even when combined, they are relatively small. As a result, perhaps the only major hardware issue is whether the network-centric era can challenge Intel's current overwhelming dominance. Today, Intel is not only the world's largest semiconductor company but also one of the most profitable. For 15 years, it has maintained a near-monopoly in the most important technical aspect of the personal computer era. Clearly, Intel's future can be seen as the vanguard of the overall prosperity of U.S. hardware companies.
In the discussions earlier about network computers, cable television, cable set-top boxes, and personal digital assistants, it was shown that many of these new products are not based on Intel microprocessors. There are both commercial and technical reasons for this, but it is clear that unless there is an urgent need to use Intel, most hardware companies will not do so. Since Intel's products are inherently general-purpose and based on older designs, there are often sufficient technical reasons to choose other microprocessors. However, the widespread desire in the business world to constrain Intel's power should not be underestimated. Therefore, the emergence of other large-scale market information technology products, in itself, will bring Intel's immense success into a broader market perspective. Even if Intel maintains its position in the personal computer market, if other sectors emerge as comparable or nearly comparable markets, a single seller dominating a major field will not be seen as a factor defining industry power. The real question is whether Intel can maintain its position in the personal computer market, which continues to expand strongly.
Intel faces two major challenges—alternative processors and Intel-compatible chips. In terms of alternative technologies, RISC microprocessor vendors have received the most attention. Chapter 4 will discuss in detail the strategies and prospects of RISC chip vendors. To summarize quickly, unless the connection with Microsoft software can be broken, the prospects of RISC chip vendors in the large-scale personal computer market are limited. Microsoft's Windows NT offers true platform independence. If it becomes the standard for desktop machines, RISC vendors may have an opportunity to compete with Intel on a chip-to-chip basis. For RISC vendors such as IBM, Digital Equipment, and HP, the challenge lies in surviving in the microprocessor industry to see whether NT or a corresponding other potential product will come to fruition. These so-called fat clients clearly have a huge market. The so-called media processors may be a more likely alternative technology threat. These chips are designed specifically to handle audio, video, telephony, and three-dimensional images, and they understand that these new multimedia applications are driving the need for greater personal computer computing power. They can be used independently in non-personal computer devices or as co-processors with Intel's central processing unit (CPU). Several vendors, including IBM, MicroUnity, NEC, Philips, Samsung, and Chromatic Research, are already manufacturing these chips, and they should be available by 1997. Even Microsoft is researching this area, recently announcing plans to license its Talisman three-dimensional image architecture to interested chip manufacturers. If these chips prove valuable, they will allow personal computer manufacturers to use less powerful, cheaper microprocessors, and/or force Intel to adopt more aggressive high-end microprocessor pricing. In this sense, the multimedia nature of the network-centric era threatens to make the general-purpose technology of the personal computer era outdated.
In contrast to these speculations about alternative architectures, the real opportunities in the so-called Intel plug-compatible manufacturer (PCM, sometimes called program-compatible module) market have existed for over 10 years. Until IBM introduced its own PowerPC chip alongside Intel's Pentium, most microprocessors in its personal computers were manufactured by IBM itself. In the mid-1980s, NEC was the dominant player in the 286 market,