Proposal to the Alfred P. Sloan Foundation from the Graduate School of International Relations and Pacific Studies and The Center for Magnetic Recording Research February 19, 1998

University of California, San Diego

Table of Contents

1. The Center for Study of the Information Storage Industry: Goals
2. Existing Research Activities and Organization
   1. The Graduate School of International Relations and Pacific Studies
   2. The Center for Magnetic Recording Research
   3. Other UCSD Linkages
3. The Research Agenda of the Center: Overview
   1. Main areas of research
   2. Principal criteria for research program
   3. Globalization, Regionalism and Public Policy (Primarily IR/PS)
      1. Globalization and Regionalism. -- The Big Three -- US, Japan, SE Asia
      2. Regulation, Public Policy, Uncertainty and Firm Strategy
      3. National Strategies and International Processes
   4. Product Development, Manufacturing and Yield (IR/PS and CMRR)
   5. Competitive Dynamics (primarily IR/PS)
   6. Field Reliability and its relation to Manufacturing (primarily CMRR)
4. The Researchers
   1. Faculty
   2. Research Staff and the Development of Expertise on the Industry
5. Dissemination
6. Center Budget
7. Appendices: IR/PS and CMRR
   1. Graduate School of International Relations and Pacific Studies
      1. Process Cost Analysis for Hard Disk Drive Manufacturing
      2. Who is Us? -- The Nationality of Production in the Hard Disk Drive Industry
      3. From Partial to Systemic Globalization: International Production Networks in the Electronics Industry.
      4. Setup Time Reduction for Electronics Assembly: Combining Simple (SMED) and Sophisticated Methods.
      5. The Economics of Yield-Driven Processes.
      6. Global Strategy and Population Level Learning in the Hard Disk Drive Industry.
      7. Sustaining Competitive Advantage in Global Industries: Technological Change and Foreign Assembly in the Hard Disk Drive Industry.
      8. The Southeast Asia System and the Hard Disk Drive Industry
      9. Making Sense : Strategy Conversations in the Disk Drive Industry
   2. The Center for Magnetic Recording Research
      1. Industrial Sponsors
      2. CMRR Faculty & Research

Proposal

Center for Study of the Information Storage Industry
University of California, San Diego

     The computer industry, which is now recognized as the central force behind a new Industrial Revolution, rests on a number of enabling technologies. Best known of these is the VLSI semiconductors, especially in the form of microprocessors. But the modern computer also depends on less well-known but equally fast-changing technologies and their associated industries. One of the most important is data storage, which includes hard disk drives, removable storage, disk drive arrays, and optical storage. The purpose of this Center is to study and aid the development of the Information Storage Industry.

     The information storage sector is one of the major subsystems of the immense computer industry. Its largest segment is hard disk drives, now approaching $35 billion in global sales, and employing about 200,000 people. Dominated by U.S. firms, with some Japanese competitors, it is nonetheless a very globalized industry. U.S. final assembly firms, led by four large survivors of a decade and a half of strenuous competition (Seagate, IBM, Quantum, and Western Digital), have more than 80% of the market. Yet almost 80 percent of the industry’s employment is outside the U.S. It is one of the most technologically dynamic industries: the capacity for a given price has improved at 50% per year for the last five years, and areal density has improved at 60 percent per year since 1991. Despite its high-tech nature, it is a very competitive industry, as the rapid changes drive intense rivalry among the firms and narrow profit margins.

     Other segments of the information storage industry include disk drive arrays, removable storage, optical storage segments, and tape drives, and are no less interesting. Disk drive arrays are a $10 billion industry (gross sales). Like the HDD segment, it is dominated by U.S. companies and is emerging as a dynamic component of information storage. The market is growing faster than HDDs per se and is quite fragmented, with about 150 companies. In contrast to the HDD and array segments of the data storage industry, Japanese companies lead in optical and removable data storage. Optical storage -- CD-ROMs and new technical advances such as digital video discs -- is a $7 billion business, with non-U.S. manufacturers holding some 90% of the market. The removable storage market, which covers a range of magnetic rigid disk drives, floppy disk drives, optical disk drives, and cards using semiconductor flash memory technology, generated $3 billion in sales revenue during 1996, and Japanese firms have captured almost 80% of this segment. Tape drives are another $3 billion per year, with about an 85 percent U.S. market share.

     The structure of the industry continues to evolve, with major upheavals every few years. Historically, the emergence of new form factors (5.25 inch, 3.5 inch) has pushed some established companies into leaving the industry and allowed new players to enter. More recently, shifts in underlying technologies such as thin films and MR heads have had similar effects. Old market segments are sometimes rejuvenated, such as the dramatic growth of Iomega with removable media, and of EMC and others with disk array subsystems. Arriving in the next 18 months are near-field recording and GMR heads, each of which has received hundreds of millions of dollars of investment before the first products ship. Thus there is continual competition not just within major segments, but among different approaches to information storage.

     The storage industry needs systematic study. It embodies the complexity of forces that drive high technology industries: globalization, international trade, basic research, manufacturing-technology interface, regulation, public policy, markets and finance. The firms in this industry face stunning managerial challenges: they must rapidly develop new technology, infuse that technology into efficient manufacturing processes that make defect free outputs at low cost, manage workforces of wide cultural range around the globe, deal with many governments having different regulatory rules, navigate the vicissitudes of international shifts in money values and trade patterns, and sell their output to final users, the computer companies, who lie between these firms and the consumers who buy the end product.

     The industry will greatly benefit from a Sloan Center. It has been successful and impressive. That success, its causes and its problems deserve study. Rapid change leaves the people in the industry little time to reflect on what they have done, how they have gotten to the present, and what might shape their future. The industry has long invested in basic technical research, both in its own firms and at Centers such as the Center for Magnetic Recording Research (CMRR) at UC San Diego and Carnegie-Mellon’s Data Storage Systems Center. It does not, however, have a Center able to connect that basic research on technology with analysis of markets, manufacturing, regulation and international trade issuesissue that also influence the industry.

     We propose to develop such an Industry Center housed at the University of California at San Diego. The purposes of the Center include:

     The Center will link together the existing research capacity of UCSD's CMRR in the technology of magnetic recording with the research capacity of the UCSD's Graduate School of International Relations and Pacific Studies in economics, management and international public policy. It will also draw on other intellectual resources at UCSD in departments of Economics, Political Science, Sociology and the School of Engineering. The Center will draw on existing linkages between industry and the university developed by the CMRR and by the existing IR/PS project on the data storage industry. It will provide for continued interaction with and outreach to industry through conferences, open seminars, papers, and a "virtual library" concept via the World Wide Web.

     Research on data storage at UCSD is concentrated in two academic units, The Graduate School of International Relations and Pacific Studies (IR/PS) and the Center for Magnetic Recording Research (CMRR). Both have active contacts and research with the information storage industry.

The Graduate School of International Relations and Pacific Studies.

     IR/PS is a professional school of international affairs, the only such institution in the University of California. Through several programs (the Masters in Pacific International Affairs, the International Career Associates Program, and a Masters in International Technology Management), it trains students for careers in international affairs, combining management, economics, international relations, public policy, the social sciences, foreign language and area studies. Founded in 1986, IR/PS has a faculty of 25 specialists in economics, management, international relations, and area studies. It currently houses two prestigious journals, the Journal of Economic Literature and International Organization in political science. Its alumni work all over the world.

     IR/PS is currently home of the Data Storage Industry Globalization Project, funded by the Sloan Foundation. Two faculty members of IR/PS, Roger Bohn and Peter Gourevitch, have been leading the Project since 1995. Participants include Dr. David McKendrick, Project Director, Professor Stephan Haggard, UCSD graduate students, and colleagues at other institutions. At present, the project concentrates on the global dimensions of hard disk drives. It examines the location decisions of industry, the spread of employment around the globe, the relationship between yield and labor costs, and the influence of national vs. international factors on the strategies of firms. The research seeks to analyze globalization through the interaction of four forces that shape firm behavior: factor costs, government regulation, agglomeration, and networking. To date we have conducted over 200 interviews with industry personnel, and hundreds of other contacts through mail and telephone surveys. Among the major findings to date are:

     See the Appendix to this proposal for a summary of research activity.

The Center for Magnetic Recording Research

     Founded in 1983 by a consortium of companies in partnership with UCSD, CMRR receives research funds from its core industrial sponsors, which are leveraged by grants from federal agencies. IBM, Seagate Technology and Quantum are three of the more than dozen current sponsors. The CMRR's research lies in the area of magnetic storage dealing with micromagnetics/recording physics, materials for media and heads, mechanics and tribology, read/write channels, and advanced instrumentation. It has four endowed chair faculty. CMRR also has a number of visiting faculty researchers and it trains students at the Ph.D. level who go on to jobs in industry and universities.

     CMRR has well established, on going contact with the major companies of the information storage industry. Its sponsors sit on an advisory board of the Center. Many Center alumni have gone on to work in t he industry, and some of its staff, such as its Associate Director are alumni from the industry itself. The Center publishes research papers which are distributed to sponsors in addition to being published in research journals. The Center sponsors many semi-annual conferences, weekly lectures and many meetings over the years which are heavily attended by people from the industryindusry.

     At CMRR, the research activities will be coordinated by: Professor Sheldon Schultz, Director of the CMRR and professor of Physics; Gordon Hughes, Associate Director of CMRR and previously an executive at Seagate; by Professor Mohan Trivedi, director of UCSD’s Program in Advanced Manufacturing; and by Professor Kenneth Kreutz-Delgado.

Other UCSD Linkages

     UCSD has other resources of value to a new Center. The School of Engineering houses the faculty affiliated with the CMRR. The School of Engineering and IR/PS together run the Program in Advanced Manufacturing for master’s students in engineering, several of whom have interned in information storage companies. The Departments of Economics and Political Science provide expertise on management and policy issues. Area studies departments provide expertise on the various countries in which the industry operates.

     The new "Center for the Study of the Information Storage Industry" would build on the existing relationships of the CMRR and IR/PS with the storage industry to extend existing research activities, explore new areas, continue human resource training, and develop industry linkages. At present, the CMRR has considerable strength in basic research on magnetic technology, but has not linked that knowledge to issues of manufacturing and the economics of the industry. Conversely, IR/PS has considerable expertise in issues of economics, management and regulation, but no expertise in the specific technologies of the industry. This separation has worked well with the distinctly different projects of the two units --- the CMRR on basic research in technology, IR/PS on globalization. But at this point, there is value to tighter collaboration. Working together will allow the linkage of technology to manufacturing issues, and manufacturing in turn to issues of economics, management, policy and trade.

     We plan that the Center will concentrate its research in four main areas. All four will be directly useful for industry participants, but in other ways they are quite different. One of them is an analysis of the ongoing interactions among globalization, regionalism, and public policy; one is a factory floor technical effort aimed at improving a specific key aspect of how drives are designed and built for high reliability; one is a combined managerial and technical look at the methods and tools of product development and manufacturing; and one examines competition within and across segments of the storage industry.

     The four areas are:

• Globalization, regulation, and regionalism. Through the globalization project we have been looking at the industry's geographic evolution, the forces behind it, and the implications for America. There are many unanswered questions, and we wish to continue the research while linking it to specific manufacturing issues. These issues tell us a lot about public policy, welfare effects of globalization, and managerial strategy.
     
• Fast product development, manufacturing ramp-up, and yields. We will compare and analyze companies’ methods for fast product development and ramp-up to high yield manufacturing. Speed is a vital aspect of this industry, and while all the survivors are quite good at rapid development, as manufacturing becomes more dispersed around the world new challenges arise. Firms will have to manage complex choices that link up quite different determinants of costs and profits: factor costs (labor prices), yields/ramp up speed, taxes and regulation.
     
• Competitive dynamics in the different segments of the storage industry. The history of the industry indicates that competition is extraordinarily fierce. Firms have found it difficult to survive technological changes. As competing technologies emerge that require a different bundle of competencies, companies will face new challenges. As storage is a part of the computer industry, it is strongly affected by changing dynamics in that industry concerning standards and market demand. For example, the recent shift of the market toward sub-$1000 computers, puts pressure on the industry to lower unit prices in addition to the historical goal of improving price/performance ratios. If Network Computers and "Net TV" machines emerge as large market segments, they will create new needs and standards in information storage.
     
• Field reliability of disk drives, to develop better understanding of failure mechanisms, and tie that back to the manufacturing and design processes. We will develop tools to exploit the new technology of drive self-monitoring called SMART, and make the crucial link from historical data about individual drives, back to the manufacturing process and drive design decisions. This part of the project makes use of a special resource at UCSD, the terabyte storage farm of our National Laboratory for Computational Science and Engineering.

     We have designed the research program around three principal criteria:

     In addition to the four main areas of research, small projects on related topics by top faculty and graduate students at UCSD will be encouraged. For example, Prof. Bohn has supervised several MS theses by participant-observers in other computer peripherals industries. One thesis developed a system for reducing setup times by 80% in surface mount printed circuit boards. Another is comparing the production of computer monitors in two sister plants in Tijuana and San Diego. These theses aid our understanding of issues important to the information storage industry, even when they are not directly in the industry. They also lay the groundwork for possible extension to other parts of the computer industry supply chain.

     We propose in the first three years of the Center to undertake research in the following four areas:

     The dynamics of this industry drive it to intense competition, rapid technological change, and vigorous globalization. It appears to be one of the most globalized of industries and one in which American firms dominate. Having done some research on these processes, we wish to continue it, and, at the same time, link it to issues of yield, manufacturing and technology management to be undertaken in different components of the project. We propose to develop in the next three years:

     a. Globalization and Regionalism. -- The Big Three -- US, Japan, SE Asia:

     Regional patterns appear strong. The industry is both globalized and yet highly concentrated. Final assembly and much production of components is located in SE Asia, with a local headquarters around Singapore; product development and research are concentrated in the US and Japan, along with some manufacturing of high value added components. Europe is a bit player, with minor parts manufacturing and some final assembly. Within these regions there are other patterns of concentration: in the US for example, much activity concentrates in California, with other centers in Colorado and Minnesota.

      What drives these patterns and concentrations? "Agglomeration" effects (the advantages of proximity) have been known for many years. Is this one of these processes? What role do countries and country policy play in this process? Can other countries enter the game? How will regionalism shape the future location of the industry and the distribution of rewards? How does this phenomena influence firm strategy and location choices?

      Several parts of our project deal with these issues:

     Describing the global system. We plan to continue describing the pattern of globalization. Our data on this has been one of the most talked about things we’ve done. Current competitive pressures in the industry have caused more changes since our 1995 benchmark: IBM ended final assembly in San Jose, and Seagate just ended final assembly at its plant in Clonmel, Ireland. This material will allow us to describe more accurately just what concentrations and what specializations have occurred in each part of the system.

     At the same time, we will supplement our work by developing another benchmark of employment by skill level in each country. Many of the arguments on globalization turn on the shift of value added activities. The U.S. claims to keep the high paid jobs in research and product development. To track this carefully, we need data that looks at the structure and composition of employment within the United States and compare it to structure and composition of sites outside the U.S.

     The SE Asia system. We have done substantial research on these questions in the SE Asia component of our project led by Professor Stephan Haggard of IR/PS, Professor Richard Donor of Emory University, and Professor Wong Poh-Kam of the University of Singapore. We wish to continue their activities and to develop some new elements of it:

• Singapore and the China Circle: The division of labor within the SE Asia region is complex. Singapore seeks to be a leader, its own type of Silicon Valley. China is becoming an important participant in the regionalization of the disk drive industry. It does so as part of a system, the China Circle, that links up China, Hong Kong, Taiwan and Singapore. The HDD industry is part of a larger division of labor in the computer industry involving these countries. The countries and leaders in them play complex roles in the supply of labor, technology, management skills, production, capital. IR/PS Professors Naughton and MacIntyre will focus on this aspect of the project working with Professor Haggard and his team.

• Trust relationships in the globalization network. Many activities in the disk drive system take place where there are no strong institutions to enforce contracts or manage a transparent regulatory process. This goes against of much conventional wisdom in the US, which says that efficient markets require clear and transparent regulations, courts and enforcement mechanisms. Recent research on "trust" argues that other mechanisms can substitute for formal ones in making markets work. These other mechanisms include ethnicity, religion or family, which substitute for formal institutions of contract enforcement and courts. Professor Rauch of our Economics Department will focus on this part of the project working with Professor Gourevitch and Dr. John Richards.

      b. Regulation, Public Policy, Uncertainty and Firm Strategy:

     In studying globalization, we have noticed important effects of regulation and public policy on firm strategy. Public policy shapes taxes, subsidies, local content laws, trade barriers, regulatory policy, financial systems, corporate requirements. These are referred to by many in the industry as major influences on location decisions and firm behavior. We wish to continue this research by analyzing carefully the role of regulatory processes in shaping the firm environment. At the same time, we wish to understand how firms evaluate and analyze regulation and uncertainty in forming their own strategic choices. Professor Gourevitch will lead this part of the project, working with Dr. John Richards.

     c. National Strategies and International Processes:

      As firms compete, do they follow patterns set by nations from which they come, or do they converge on practices shaped by international market forces? Dr. McKendrick has been analyzing this and will continue to do so. Invention and technology have generally come from the U.S., notably IBM. Yet some Japanese firms also consistently have very good technology, without having much market share to show for it. Neither the U.S. nor Japan is ending up as the center of manufacturing, despite the fact that most demand comes from American computer companies.

II. Product Development, Manufacturing and Yield (IR/PS and CMRR)

     The second major research area of the Center will look at speed and effectiveness in product development, and in bringing products into successful production. Speed is a vital aspect of this industry, as in most high technology. Technological progress is among the most rapid in any industry (60 percent per year improvements in areal density, 50 percent per year reduction in cost per megabyte, product lifecycles from 9 to 18 months), and a difference of a month in getting a new product into high volume production makes a tremendous difference to its life cycle profit. Presently the industry brings new drives into production at low yields (70 percent) in order to capture high initial prices and to rapidly build manufacturing experience. But as output of a drive ramps up to 10,000 per day it is economically and physically impossible to continue to rework 30 percent of the drives. Yields therefore must and do improve rapidly in most ramp-ups.

     Of course, all the survivors in the industry are quite good at rapid development. But the continued pressure to bring new technology to market rapidly leads to continual challenges. And as manufacturing becomes more dispersed around the world new issues arise.

     We will conduct the research in two closely coordinated parts. One will look at manufacturing in assembly and media plants, emphasizing yields, and the speed with which new products are "ramped-up" to high yields and output levels. These activities take place primarily outside the U.S. The other will look at what happens before manufacturing ramp-up, namely the product development process and the transfer from development into manufacturing in a different country. The two are vitally linked, as good development and good transfer mean that fewer problems have to be solved during ramp-up.

     Following the example of the Berkeley semiconductor project’s research on manufacturing, this research will be conducted jointly by faculty and students from IRPS, CMRR, and other parts of UCSD. The objectives of the proposed research are as follows.

• Benchmarking: First, we intend to collect detailed data on different measures of product/process development as well as manufacturingfinal disk drive assembly. Previous studies (e.g. the MIT International Motor Vehicle Program, Berkeley’s semiconductor study. Leachman, Excellence in Electronics) have shown that such data can be of direct use for participating companies. The data is especially valuable when compared across companies, which allows identification of best performance in the industry and can be used for benchmarking within the industry. We will be answering questions such as "How fast do different sized development projects go? What are initial yield levels when transferring designs of a given complexity and novelty? How fast can yields be increased? Are some factories consistently better than others at ramping up new products?"

• Performance drivers: Second, this data collection effort will put us in a position to identify and study a number of performance drivers, such as how project teams are organized and coordinated internationally, and how people are shifted around the world to help with the transfer of new products into manufacturing. As shown in the IMVP studies, most notably Clark and Fujimoto 1991, such an analysis can reveal fundamental patterns that distinguish high performing organizations from low performing organizations. Typically, these patterns cannot be detected by traditional single case studies. Key variables can include aspects of project organization, development tools, transfer processes, manufacturing methods, and many more. Furthermore, we could collect detailed data on the transfer between development (U.S.) and manufacturing (Asia) at the level of the individual development project.

• Modeling and implications: This research will allow us to better understand yield-driven manufacturing processes, the relationships between development and manufacturing, and the nature of product development in industries where there are many development projects in quick succession, along a relatively predictable technological trajectory. In contrast to autos and semiconductors, development projects in this industry are much smaller and faster, providing a useful contrast. The research will help also us understand the options and likely future of the global locations patterns in the industry. For example as Singapore attempts to import some R&D work, will it be practical to have process development done in Singapore, product development done in the U.S., and manufacturing done in a third country? Or, do product and process development have to remain co-located, so that both are done in the U.S. or both are done in Singapore?

Contribution to Industry: This research on product development and yield allows us to identify best practice, and the range of practice among successful companies; develop a number of key performance measures that can serve as a basis for an ongoing benchmarking in the industry (include product measures, e.g. parts commonality; and process measures, e.g. ramp-speed); isolate determinants of performance; and get new insights for faster development and ramp-up. These insights will help companies do better in future product development and manufacturing.

Contribution to Research: Little is known on the transfer and ramp-up processes. The integration between product and process development seems to be key in this industry, whereas most previous studies about design/manufacturing issues have looked at traditional design for manufacturability. From a globalization perspective we are interested in (a) the transfer, i.e. how does the U.S. site prepare the transfer, how is the transfer managed, and what are its consequences; (b) to identify regional differences in where things are done, and especially where things are done well. On product development, while several previous cross-firm studies of the product development process have been done, notably in autos, the short life-cycle of HDD development gives us the opportunity to study several complete development cycles in each firm, and to tie development to transfer and ramp-up.

III. Competitive Dynamics (primarily IR/PS)

     While the Data Storage Industry Globalization Project has concentrated on globalization issues, a Center would deepen its understanding of and relevance to the storage industry through a more general analysis of the competitive dynamics within and across the hard disk drive, optical storage, disk drive array, tape, and removable storage segments of the industry. Some technologies underlying the different segments are complementary, some are in competition. As the industry as a whole continues to evolve, a number of issues deserve scrutiny. Following are some examples of interesting research topics in this area.

     1. In one sense, the history of the disk drive industry is a story about companies on a treadmill. Firms are under relentless pressure to introduce new products with ever greater capacity before the competition does. We have found that many formidable companies were unable to stand the pace: Hewlett-Packard, Burroughs/Unisys, Siemens, Digital Equipment, Sony, and Texas Instruments, just to name a few of the more than 125 firms to have failed in this industry. Although speed is a critical success factor in hard disk drives, there is anecdotal evidence that firms can run too fast. That is, many companies that led the industry with the most innovative drives (i.e., high capacity) found it difficult to find a sufficiently large market and exited the industry. For example, a host of U.S. companies -- Micropolis, Atasi Corporation, Ibis, Applied Information Memories -- embraced such a "high capacity" strategy, which is commonly associated with the American style of innovation. The core idea is to innovate your way out of the low margin, high volume segments populated by firms whose singular capability is inexpensive manufacturing. Instead, the decision to avoid the high volume, intensely competitive segments may have actually weakened the firms that adopted this strategy. Little is known about the conditions under which being a market leader or follower is a viable strategy. Is it necessary to have a presence in high volume segments in order to survive? Can a firm be late to a capacity segment? As time passes, do large companies become less subject to failure? Do the benefits of being among the capacity leaders depend on the intensity of competition in those segments? This project would analyze systematically the relationship among entering new capacity points, growth and firm survival -- the pace and path of strategic change. Such findings may prove useful to firms in emerging storage segments.

     2. Disk drive companies have always faced threats from competing technologies, whether real (the desktop disk drive) or imagined (magnetic bubbles). Sometimes the challenges emerge in an evolutionary way. For example, during the 1950s and 1960s, magnetic drums and head-per-track disk drives competed with HDDs employing flying heads, and companies in each segment were well aware of each other, as well as the advantages and disadvantages of the competing technologies. Other times challenges come in punctuated shocks, when change occurs very rapidly. Clayton Christensen has shown that the introduction of a new, smaller form factor challenged incumbent companies relatively quickly. Incumbents were unsuccessful, he argues, because they were embedded in a particular "value network" and were unable to see the possibilities inherent in an emerging value network. That is, companies selling disk drives to mainframe manufacturers did not discern the needs of other value networks such as the minicomputer market; those making HDD for minicomputers were unwilling to commit resources to develop disk drives for desktop computers, etc.

     However, Christensen never tested this thesis statistically, and alternative explanations are possible. For instance, most disk drive firms failed before they entered a second form factor; many of these also failed before any other firm introduced a new form factor. This suggests that, for the average firm, competition is most intense within form factors. At the same time, one needs to specify more carefully the characteristics of the firm that give it a higher probability of surviving the movement into these new kinds of technologies. One is the length of time a firm has been in the industry. Organizational fates are often traced to causes at the time a firm entered an industry. Such entry conditions specify the routines and capabilities needed to survive. New technology may redefine these capabilities, and well-developed organizations may suddenly be poorly fit. In other words, the length of time a firm has been in the industry may be a better predictor of survival than the kind of customers it serves. One could make the case, for example, that Seagate, Maxtor and Quantum brought such new skills and, as specialists, were better organized to compete in the new environment than Burroughs, Control Data, and Digital Equipment. Similarly, size may enable firms to survive the transition to new technologies even if they are otherwise weak competitors. This project proposes to examine the effect of new form factors in the HDD industry on firm survival. Are innovations brought by existing organizations or through the founding of new organizations? Are companies more likely to fail when moving from one form factor to another than they are competing within a given form factor?

     3. Firms attempt to leverage their knowledge and extend survival by diversifying into related technological fields. As the hard disk drive industry matures, companies may find it necessary to diversify into new markets. Seagate, for example, has diversified into tape drives and software. But when an organization operates in more than one market it is subject to conflicting competitive pressures. What are the conditions under which technological diversification succeeds? This project would examine two kinds of technological change within the storage industry. One is the diversification of companies from floppy disk drives into hard disk drives. In the past, a number of floppy disk drive makers diversified into hard disk drives. Some were able to derive economies of scope out of their R&D resources: Iomega, Micropolis, Shugart Associates, and Tandon. Others tried and quickly failed: Sony, Qume, Micro Peripherals, Olivetti, and PerSci. Why were some FDD firms able to make the transition to HDD? What were the characteristics of the firms that entered HDD but exited soon thereafter? The benefit of this kind of historical study is that the amount of information available over a long period of time may yield some conclusions that prove useful to firms considering similar diversification moves now.

     The next project would explore an ongoing technological diversification: the movement of hard disk drive companies into disk drive arrays. The array market is emergent, with some 150 companies, and consists of some complicated patterns of competition. Some companies that failed in the disk drive industry have become quite successful in designing and assembling disk drive arrays: Storage Technology, Digital Equipment, Data General, Control Data and Hewlett-Packard. Others are computer systems manufacturers such as Compaq, IBM, Tandem Computers, Sun Microsystems and AST Research. Still others conform to what we have observed in the HDD industry -- a host of companies from other industries, including EMC, the industry leader. At this early stage in its development, the array industry resembles the evolution of HDD industry in another way. The captive producers like IBM still control some 60% of the market. Will array specialists outcompete the captives, as occurred in disk drives? Why have some HDD firms found success in arrays? Why are de novo startups less prevalent in arrays than they have been in HDD?

     Computer disk drive performance has advanced remarkably, in reliability as well as capacity and speed. Sophisticated error control algorithms are designed into all drives, using electromagnetic tests and digital coding techniques to detect and correct data readback errors. Of course massive failures covering whole subsystems cannot be recovered from by these means. Therefore the error detection algorithms are being extended to predict drive reliability problems in advance of actual failures.

     This predictive reliability technology, termed "SMART" (Self Monitoring and Reporting Technology), is now an industry standard for monitoring high-end drive performance and reporting incipient failures to the operating system. This allows the drive’s user to take action before actual failure occurs. However, current SMART techniques are only partially effective in predicting failures early enough to be useful.

     UCSD and its National Laboratory for Computational Science and Engineering (formerly the Supercomputer Center) have initiated a study to improve SMART accuracy. The study has unique capabilities, beyond what any disk drive company can do on its own. It provides a controlled, statistically valid experiment, using advanced data analysis techniques on several thousand drives at a cooperative site, namely the National Laboratory’s terabyte data farm. IBM is the lead drive manufacturer working with UCSD to conduct this experiment. Seagate Technology drives are planned to be added, and the disk drive industry standards organization (IDEMA) has also agreed to participate. The study’s goal is to assist drive manufacturers in improving SMART technology to reach 90% accuracy in predicting drive failures one day in advance.

     As part of the new Center, this project will be extended to look at the drive manufacturing process. Drive manufacturers gather large amounts of reliability test data on new drive designs, but time to market competition requires that drives go into production with only months of statistically valid reliability data. However, five year life guarantees are common. This contradiction is an industry-wide concern, which this project will address. Our goal is to develop the ability to reject drives which are candidates for failure within five years, and to provide corrective feedback to drive designs.

     Background: Competitive pressures for speed force drive makers into concurrent development of designs, manufacturing processes, component technology, reliability testing, and customer qualification. This leads to product readiness risks. Manufacturing volume ramps can and do stall, and product marketing windows are missed, when unexpected R&D is needed to solve unpredicted manufacturing or reliability problems. These are a significant cause of the large profit swings that the industry is known for.

     We propose to extend SMART from detecting incipient failures one day ahead in the field, to using it during final testing inside the manufacturing plant. The concept is to use it during test to validly predict five year reliability, by detecting symptoms of off-spec conditions that are known to cause reliability problems later. This would allow considerable reduction in the risks incurred by concurrent development.

     In addition, this extended SMART technology offers an opportunity to provide predictive feedback on latent manufacturing technology problems during pilot production when the product design and manufacturing process are in final development. The goal is to learn how to use SMART predictions of problems to more quickly detect latent manufacturing and design problems on prototype and pilot drives. This would allow more of these problems to be solved during product design, and thereby reduce the expense and waste of transferring an unstable drive design into production.

     SMART technology offers an opportunity to provide predictive feedback on these latent manufacturing technology problems during the pilot production stage, when the product design and manufacturing process are in final development. Without SMART, reliability engineering requires time-consuming testing to try to force drive failures. The goal is to learn how to use SMART predictions of incipient problems to more quickly detect latent manufacturing and design problems in prototype drives. This could allow these problems to be solved during product design, and thereby reduce the expense and manufacturing resource waste of transferring an unready drive design to production.

     There are two critical milestones for succeeding in this extension of SMART technology to prototyping and production ramp management. First, the ongoing program has to succeed in its goal of increasing the predictive accuracy of SMART. Second, one or more disk drive manufacturers must be willing to participate in the extension, and open their manufacturing technology for study. Although this involves industrially sensitive areas, the semiconductor industry has succeeded in similar studies, such as the well-known Berkeley semiconductor project sponsored by the Sloan Foundation. The cooperation so far from IBM and others argues that the extension will be possible.

     The various research activities described above will involve regular faculty, post doctoral research associates and graduate students.

     Roger Bohn and Peter Gourevitch will continue their active role as we shift from Project to Center. In addition, we plan active involvement of other faculty.

     At CMRR, the SMART project involves a number of engineering faculty, including Professor Mohan Trevedi, Ken Kreutz-Delgado and others. Gordon Hughes, Associate Director of CMRR and previously an executive at Seagate, will be the project manager for the SMART portion of the Center. UCSD’s Program in Advanced Manufacturing, directed by Prof. Trivedi and affiliated with CMRR, is seeking as well to obtain permission to recruit a new faculty member in the field of manufacturing engineering. This person will be active in the research on rapid ramp-up and product development. The Sloan Grant will assist in this process by showing external support for these activities.

     From IR/PS and the Social Science Departments at UCSD, we plan the involvement of several faculty. Professor Stephan Haggard will continue his role as our leading associate dealing with the SE Asia portion of the project. He will continue to work with Professors Rick Doner of Emory University, and Professor Wong Poh- Kam of the University of Singapore. Professor Barry Naughton, economist, specialist on China, researcher on "the China Circle", the regional economy involving the several Chinas. Prof. Andrew MacIntyre, political scientist, specialist on Indonesia and SE Asian political economy. Prof. James Rauch, professor of economics, UCSD, specialist on international trade.

      We also have had useful conversations with other faculty who have agreed to continue to work with us on an informal basis: Professors Takeo Hoshi, Eusyong Kim, Mikhail Klimenko, Lawrence Krause, John McMillan, and Chris Woodruff all have expertise relevant to our activities.

     As noted above, the School of Engineering hopes to hire a new faculty member that would be able to contribute to this project. IR/PS may also be able to deepen its participation through two faculty searches underway in the field of technology management.

     As part of Sloan’s goals in developing human resourceresourse skills, experts in industry studies who can go on for careers in research and managementmanagment, we will continue to employ people at various levels of their education. We have found very effective the mixing of post-doctoral fellows, current Ph.D. candidates, and professional masters students at IR/PS and the PAM degree program.

• Post doctoral fellows. Recent Ph.D’s can provide essential assistance to our research and help the Sloan foundation achieve its goals of training a cohort of industry experts. Post doctoral fellows have developed the research skills needed to do this kind of work. They also have incentives to work with us and post doctoral positions are a way of building their research skills. They can build the record and diversify their knowledge that helps them move forward, and at the same time, develop the industry specific contacts and network about industry specific work that allows them to contribute to industries later in their careersgo forward. This year we added two post-docs, Christian Terwiesch, recent Ph.D. from INSEAD in Europe, and John Richards, recent Ph.D. from UCSD. They have been quite valuable and we expect to continue their involvement in the Center after they leave to take full time faculty positions at other universities.

• Ph.D. candidates. Ph.D students are effective participants in the research process and can be influenced to develop industry skills. We have worked effectively with Allen Hicken in this regard, and expect to continue him next year. We have a few other possibilities in the pipeline as well. The balance between Ph.D’s and post-docs depends on the supply of each. If we have strong and available Ph’D.s we can use them; if they are scarce, and good post-docs are available, we will employ them. Both contribute to human resource skill development.

• Professional masters students. The Masters students have been helpful as well and we plan to continue to use them. They come from the both IR/PS MPIA program (Masters of Pacific International Affairs, a two year professional degree combining managementmanagment, economics, and internationalinternatiuonal public policy) and the Program in Advanced Manufacturing. These students are able to do certain kinds of research and working with us helps train them for industry jobs. A number of them have already taken jobs in the information storage industry.

     Through using post doctoral students, Ph.D. candidates, and Masters students we are able to contribute to Sloan’s objectives in developing new talent, both for industry and as future faculty.

     We seek dissemination of our ideas, both to scholarly and industry outlets.

Public Dissemination

     We have published articles in the two most widely read storage industry trade journals:

Peter Gourevitch, Roger Bohn, and David McKendrick, "From San Jose to Singapore: Global Portrait of an Industry." Data Storage, Vol. 33, No.2, July/August, 1997.

Roger Bohn, Peter Gourevitch and David McKendrick, "Globalization of Disk Drive Manufacturing: What and Where." Insight, Vol. 10, No. 4, July/August, 1997.

     Data Storage received such positive feedback from its readers that it has asked us to contribute another article in the coming months that tracks changes in employment and location since our last contribution.

     We have also made presentations at various meetings of the HDD industry and are planning a large public conference in the Spring with CMRR at which significant members of the industry will attend. We also will conduct "outreach seminars" in Southeast Asia, to reach manufacturing people and to get feedback and contacts on our globalization research.

     CMRR runs a regular weekly seminar with a mixture of academic and industry speakers and attendees. In the past, the topics have been 80% technical and 20% managerial. The Center will use this as a regular outlet for informing industry and colleagues at UCSD about our findings.

     We plan to find other ways of working with the HDD industry to disseminate our work to them. We expect that the next wave of research will be of particular interest to the industry.

Web Pages

     Traditional publication methods are good for reaching some audiences, but are slow and don’t collect all the information in an easily accessible format. Books have extremely long lead times. Therefore, other Centers are making use of the World Wide Web as a means of making comprehensive information available quickly. The CMRR is already planning to use the World Wide Web more extensively to create a "virtual library" for its member companies. People in the companies will be able to request fax or electronic copies of articles from journals which their company libraries don’t stock. The Center’s information will therefore be visible in a "high traffic area" for the industry.

Short Courses and Seminars:

     Specialized programs for industry people provide another effective vehicle for communicating our research. We plan to work with our campus’ Extended Studies division in developing short programs. That unit has very substantial experience of this kind of work, through programs in executive education and its internationally recognized CONNECT program, which links together industry, services (law, finance, accounting) with researchers.

Research

     We have written and disseminated several working papers, and are seeking to publish them in various academic journals. We have interest from publishers for a book project. As the project grows into a Center, we seek to increase the range of activities and production. See Appendix for some of thelist of publications.

     We propose the Center be a joint project at UCSD of CMRR and IR/PS. A governing council will consist of the PIs from IR/PS and CMRR, with a UCSD Professor as the Director. As the Center evolves from its origins as a project, it will require a Manager, someone like Donna Carty from the MIT Auto Project.

Funding:

     Funds for the Center will initially come principally from the Sloan Foundation. The CMRR receives most of its funding from industry, as well as funds from the university for its faculty. IR/PS receives funding from the uUniversity for faculty support. CMRR and IR/PS will both contribute resources in faculty time, management, and space to the project.

     The budget uses a "Program Budget" approach rather than line item activity approach. That is, we focus on each area of the center’s activities and allocate funds for the goals, rather than showing the composition of salaries, travel, meetings, etc. Within each of the research activities, the line items will be a combination of faculty salary (summer ninths), post-doctoral researchers, graduate assistants, and travel.

     On the administrative side, we include activities for outreach and public dissemination, as well as administration. We envision hiring a Center Manager -- an individual who focuses not only on the administrative coordination internally, but who does extensive company relations, dissemination, and outreach.

Graduate School of International Relations and Pacific Studies

     The University of California's Graduate School of International Relations and Pacific Studies (IR/PS) is unique among U.S. graduate schools in its focus on the economic, political, and business relations of the Pacific Rim, encompassing both the Americas and the Asia-Pacific region. IR/PS is recognized internationally for its excellence and attracts students from every region of the Pacific Rim and beyond. Transitions Abroad has ranked IR/PS among the world's top 10 international graduate schools.

     As the only professional school of international affairs in the University of California system, the school's primary objectives are:

• to prepare students with an interest in the Pacific Rim for positions of leadership in business, government, diplomacy, policy analysis, journalism, and other fields;



• to serve as a center of excellence for research on economic, political, social, technological and security issues confronting nations of the Pacific Rim; and



• to promote international dialogue on issues involving the Pacific Rim.

     The IR/PS faculty is internationally renowned for its expertise on the Pacific Basin. Our faculty members earned their doctoral degrees and built their reputations at many of the top institutions of higher learning in the United States and abroad. The faculty span both business and political science fields. Scholarly investigation of mixed policy/management issues form an integral part of its core mission. Moreover, scholarship on East Asia is being augmented by growing faculty strength in Southeast Asia.

     This faculty and the School form the home for the Sloan Data Storage Industry Globalization Project, which beganwas begin in July, 1995. Since that time the principal investagors, the project director and the researchers have developed an in-depth knowledge of management and public policy issues associated with the hard disk drive industry. The project has produced various working papers, and expects to generate more:

     Designers, engineers and manufacturing managers in the hard disk industry need tools that allow them to predict how different parts of their manufacturing systems affect the costs of their products. Current methods for doing cost analysis are inadequate. Both traditional methods and activity based costing fail to capture the inter-relationships between manufacturing processes. This paper proposes a process based system, Process Cost Analysis (PCA), a system that combines engineering models of the manufacturing system with accounting figures and records to provide managers with information about their manufacturing costs, the ways in which those costs arise, and the costs of alternative manufacturing methods. This paper discusses: (i) Several arguments for the necessity of using process modeling to understand costs in manufacturing; (ii) A process model based method, Process Cost Analysis, to clearly associate the processes which cause costs with the allocated costs of manufacturing; (iii) Yield factors, a method for simplifying the representation of complex processes to allow localized studies of global yield effects in complex manufacturing processes; (iv) Two illustrative examples, one of which uses studies conducted over a six month involvement with Maxtor, Inc.

     The purpose of this research was to provide an empirical snapshot of globalization in the hard disk drive industry. This revealed that globalization is a complex phenomenon, and the location of the HDD industry varies dramatically depending upon the measure being used. we found that if the measure is: (1) nationality of the firm that does the final assembly of HDDs, then the U.S. has 85% of world production, Japan about 15%; (2) the geographic location of final assembly activity by country, then the U.S. share in 1995 was 5% and Southeast Asia held 64% of world production; (3) the location of employment in the industry, including R&D and the various stages of production, then the U.S. has about 20% of world employment, larger than any other country but decidedly not the bulk of world employment; and (4) the location of wages paid in each country, which reflects the actual amount of wages paid, then the U.S. share of total wages is about 42%, substantially larger than its share of total world employment. These findings suggest policy makers should be cautious when designing policies to address the impact of "globalization."

     This paper attempts to provide a conceptual framework for thinking about "globalization" in the data storage industry. International production has expanded rapidly. In order to cope with the increasingly demanding requirements of global competition, companies are forced to integrate their stand-alone operations in individual host countries into increasingly complex international production networks. This paper argues that globalization results from firms searching for clusters of specialized capabilities and accessacccess to contested growth markets. As a consequence, globalization is more complex than in the past: (1) new products are produced overseas much earlier than predicted by the product life cycle theory; (2) the scope of international production has been substantially broadened to include high value-added activities; (3) outsourcing also includes more high value-added functionsfinctions; and (4) systemic rationalization now cuts across national borders and covers a variety of cross-border linkages.

     Setups determine downtime, capacity, product quality, and to some extent costs. As much as 50% of effective capacity can be lost to setups in some electronics assembly. This paper shows that radical reductions in setup time are possible for electronics assembly as they are in other industries. The paper uses a two-part approach. The first part consists of classic process re-engineering using "Single Minute Exchange of Dies" (SMED) concepts developed by S. Shingo for metal fabrication. The second part uses a sophisticated factory information system, with hand-held wireless computers and barcode scanners, to further reduce setup times and increase setup accuracy. This two-part approach gave a reduction of about 86% in key setup times, plus labor savings, quality improvements, and other benefits. We estimate that the setup reduction had a payback period of two months. The results validate the SMED concept, and show that it is applicable well outside the traditional domain of die setups for stamping machines.

The economic performance of many modern production processes is substantially influenced by process yields. Their first effect is on product cost -- in some cases low yields can cause costs to double or worse. Yet measuring costs alone can substantially underestimate the importance of yield improvement. We show that yields are especially important in periods of constrained capacity such as new product ramp-up. Our analysis is illustrated with numerical examples taken from hard disk-drive manufacturing. A one percentage point increase in yields can be worth \$2 million per month, about 6 percent of gross revenue and 17 percent of contribution. In fact an eight percentage point improvement in process yields can outweigh a \$20 per hour increase in direct labor wages. We interpret this to mean that yields, in addition to or instead of labor costs, should be a focus of attention when making decisions such as new factory siting and type of automation. The paper also works out relative effects of first pass yield and rework yield.

     Some scholars argue that firms within the same industry engage in similar foreign investment behavior irrespective of nationality because they face a common set of pressures and incentives. Others emphasize the persistent diversity in business practices and global behavior of firms from different nations. Using the hard disk drive industry as a case, this paper explores whether nationality influences how firms in the same industry globalize and, if so, whether it affects industry performance. The findings suggest that firms from the same nation are likely to adopt similar global strategies initially, but that, over time, the industry as a whole converges on the same blueprint for action. We also find that the national industry that is first to select what becomes the dominant strategy acquires an advantage over competitors from other countries. That is, the timing and direction of globalization matter for a national industry.

     One of the interesting facts about the hard disk drive industry is that American firms have been the dominant competitors since the industry’s inception. The purpose of this paper is to examine why U.S. firms have been so resilient. Traditional explanations for industrial competitiveness include home market demand, form of industrial organization, innovation, and the role of private and public institutions. While these can be central to the development of an emerging industry, they are insufficient for sustaining competitiveness once an industry matures. This paper argues that the globalization of production can also be an important complementary factor in sustaining the competitive advantage of national industry. By moving quickly to secure the benefits associated with overseas production, national industry can maintain its resilience in the face of foreign competition.

     This is a general name for a set of papers being written that explore in depth the division of labor among countries in Southeast Asia as they manage production in the disk drive industry. These papers have not as yet been printed as working papers in the Project.

     Managing and making policy in this industry faces formidable challenges to information, as technology and markets change so rapidly. Managers make use of "information" shorthands to simplify their choice options. Understanding these shorthands helps us understand the "conversations" among managers which structure choices they face and guide their strategic judgments. This allows us to see what points get emphasized and what points get underplayed or neglected.

     The Center for Magnetic Recording Research (CMRR) was founded in 1983 by a consortium of the U.S. magnetic recording industry to perform research in magnetic disk and tape storage. The initial funding was used for the construction of the CMRR building, for research equipment, research funding, the Information Center, and the endowment of four chaired professorships. At the outset, the Regents of the University of California provided the land and additional funds for the Center building. The current Industrial Sponsors provide part of the research expenses, which are leveraged by grants from federal agencies, and the commitment of the UCSD campus for funding of faculty salaries and the maintenance and utility expenses of the CMRR building.

     The Center is housed in a 26,700 square foot building on the campus of the University of California, San Diego. The modular, three story research facility provides space for laboratories, academic and administrative offices and support areas. It includes a 100-seat lecture hall, two conference rooms, and an Information Center library dedicated to all aspects of magnetic storage.

     The Center's mission is to advance the state-of-the-art in magnetic storage, and to produce highly trained graduate students and postdoctoral professionals. CMRR also serves as a catalyst for joint investigations amongst its industrial sponsors, government agencies, and the University.

     The Center has a core group of five faculty in residence, from the departments of Electrical and Computer Engineering, Mechanical Engineering, and Physics. There are additional faculty from UCSD, UCB, SDSU, and several other academic institutions, supported by CMRR related programs. The combined Center activities have over thirty graduate students in residence, typically seven postdocs, several senior research associates, two electronics technicians, and a modest administrative and clerical staff.

     The Industrial Sponsors provide advice and direction to our programs through an Advisory Council which meets at the twice yearly Research Reviews. Its current Industrial Sponsors include:

* Eastman Kodak

* GEC Plessey Semiconductors, Inc

* Hewlett-Packard Laboratories

* Imation Corporation

* International Business Machines (IBM)

* Lucent Technologies

* Maxtor Corporation

* National Security Agency

* Overland Data Inc.

* Phase Metrics, Incorporated

* Quantum Corporation

* Read-Rite Corporation

* Seagate Technology

* SGS-Thomson Microelectronics

* Storage Technology Corporation

* Western Digital

     CMRR is an active member of the National Storage Industry Consortium (NSIC), and has been a recipient of significant sub-contracts from their ATP and ARPA programs. As envisioned in the initial planning of the Center, the faculty have developed their diverse research groups and have obtained funding from federal, state, and industrial sources, including an NSF sponsored Materials Research Science and Engineering Center.

     CMRR’s program is a collaborative approach by an interdisciplinary group of researchers to develop significant advances in ultra-high density storage, and ultra-high data rates, for tape and disk magnetic recording systems. Current broad areas include:

     Micromagnetics/Recording Physics

     Materials for Media and Heads

     Mechanics and Tribology

     Read/Write Channels and ECC

     Advanced Instrumentation

1. Professor Sheldon Schultz and CMRR Director

     Professor Schultz received the Ph.D. in Physics from Columbia University in 1960, and an M.E. degree from Stevens Institute of Technology in 1954. He came to San Diego in 1960, as part of the first group that founded the physics department, and initiated the planning for what was to become the current UCSD campus. He became active with the CMRR at its earliest stages, serving on the initial planning, search, and building advisory committees. He became an Associated Faculty Member of the Center in 1983, and initiated a research program devoted to providing a quantitative experimental set of data to fully characterize the micromagnetics and hysteresis in thin films and sub-micron particles. Prior to becoming associated with the Center he has worked in several diverse areas of condensed matter physics including Electron Spin Resonance applied to metals, dilute magnetic moments, and spin glasses; surface magnetism; weak and strong localization; and superconductivity. He has over 200 publications and has received research support from a broad group of federal agencies (NSF, AFOSR, ONR, DOE, ARPA, etc.) and industrial companies. He has served as the Director of the CMRR since November 1990.

     Professor Schultz does experimental investigation of the micromagnetics of individual and interacting sub-micron particles prepared via nanolithography, chemical processes, and electro-pore plating. He is involved in the development of new instrumentation and techniques such as MAFM (magnetic atomic force microscopy) and Kerr-SNOM (scanning near field optical microscopy) capable of measuring the magnetostatic and dynamic properties of individual sub-micron magnetic particles.

2. Professor Ami E. Berkowitz (emeritus)

     Prof. Berkowitz received his Ph.D. in physics from the University of Pennsylvania in 1953. He was employed by the Franklin Institute Laboratories in Philadelphia from 1947 to 1960 where he was manager of the Magnetism and Thermoelectricity Division. He was with IBM from 1960 to 1968 initially at the Research Laboratory at Yorktown Heights, NY, and then as manager of the Materials Branch at the Components Division in Burlington, VT. From 1968 to 1986, Dr. Berkowitz was at the General Electric R&D Center in Schenectady, NY. There he was active in electron beam addressed memories, thermomagnetic recording, magnetic tagging, magnetic printing, magnetic surface phenomena, amorphous and crystalline particles, and the development of spark erosion method for particle production. In 1986, he came to UCSD as Chair Professor in the Physics Department, and to CMRR to establish a materials group. Dr. Berkowitz has published extensively in the scientific and patent literature and with E. Kneller co-edited "Magnetism and Metallurgy", published by Academic Press. He has also given many invited talks, and has been twice appointed a Distinguished Lecturer for the Magnetics Society of the IEEE.

     Dr. Berkowitz's research is based on thin film and fine particle phenomena related to magnetic recording. It involves production of these materials and characterization of their microstructural, magnetic, and transport properties. Some current research activities include: a) giant magnetoresistance and magneto-impedance; b) antiferromagnetic oxides for MR head biasing, c) magnetic viscosity phenomena; d) new particulate and film media; e) finite size effects; and f) magnetic surface behavior.

3. Professor H. Neal Bertram

     Dr. Bertram received his B.A. from Reed College, Portland, OR in 1963 and his Ph.D. at Harvard University 1968. From 1968 to 1985 he was employed by the Ampex Corporation where he worked on fundamental problems in magnetic recording. While at Ampex he was also engaged in both theoretical and experimental studies of particle interactions, ac bias recording, media noise, the recording process for particulate and thin film media, high density and multi-track head design, and head saturation. From 1978 to 1985 he was manager of the recording physics section of the research department (ATD) at Ampex and in 1984 became a Principal Engineer. In 1985 he joined UCSD as an Endowed Chair Professor in the Electrical Engineering and Computer Sciences Department associated with the newly created Center for Magnetic Recording Research. At UCSD he has created a graduate course in magnetic recording which includes two quarters on the theory of recording materials and the recording process and a third on magnetic recording measurements. In 1986, Dr. Bertram was a IEEE Distinguished Lecturer, and in 1987 he was named an IEEE Fellow. He has recently published a book entitled: "Theory of Magnetic Recording" (Cambridge University Press, March 1994).

     Dr. Bertram's current research with graduate students involves modeling the physics of the magnetic recording process. Included are theoretical and experimental studies of high density recording for all types of advanced media: particulate, and thin film. Research is focused on understanding high density non-linearities, overwrite and erasure of the recording process, including edge track phenomena. Dr. Bertram is also concerned with noise mechanisms in magnetic media including particulate, modulation, and transition or phase noise. Since the ultimate understanding of recording involves the magnetism of the medium grains, studies of micromagnetics of particle switching, as well as intergranular magnetostatic interactions, are being performed. Analysis of magnetic thermal instabilities for single and interacting granular materials are investigated. Magnetic transducers are being studied, focusing on record head saturation and high frequency field rise time and linear signal analysis as well as pulse asymmetry and distortion of AMR and Spin Valve playback heads.

4. Professor Frank E. Talke

     Professor Talke received a Diplom-Ing. degree from the University of Stuttgart, Germany in 1965 and M.Sc. and Ph.D. degrees from the University of California, Berkeley in 1966 and 1968, respectively. From 1969 through 1986, he was working at the IBM Research and Development Laboratories in San Jose, CA, and in 1984 he spent one year as a guest professor at the University of California, Berkeley. He is the author of more than 120 publications and holds 11 US patents. His research interests are in the areas of the head/disk and head/tape interface, tribology, mechanical design related to magnetic recording technology, and high precision instrumentation. Dr. Talke accepted a position as an Endowed Chair Professor at CMRR effective March, 1986.

     Dr. Talke’s research includes a) modeling and experimental investigations of the head/disk interface of thin film magnetic media; b) investigations of the head/tape interface for longitudinal and helical scan head/tape interface, including tape surface roughness effects; c) investigation of stiction and dynamics of the head/disk interface using laser Doppler Instrumentation; d) non-repeatable run-out of ball bearings and fluid spindles; e) surface roughness studies and the relationship between roughness and wear; f) modeling of helical and longitudinal head/tape interfaces; g) interferometric studies of the head/tape and head/disk interface.

5. Professor Jack Keil Wolf

     Professor Wolf received his undergraduate degree from the Moore School of Electrical Engineering at the University of Pennsylvania and his graduate degrees from Princeton University. He has been teaching full time for over 30 years with part time stints at Bell Laboratories and the Linkabit Corporation. Prior to joining UCSD he was a Professor and former department chair of the Electrical and Computer Engineering Department at the University of Massachusetts, Amherst. Prior to that he held a faculty appointment at the Polytechnic Institute of Brooklyn. Jack is a Member of the National Academy of Engineering and is a former NSF and Guggenheim Fellow. He was the first faculty member appointed to CMRR. He is also a part time employee of Qualcomm Inc., San Diego.

     Professor Wolf studies many of the topics normally associated with the design of communication systems since the recording of digital data on magnetic media has many similarities to the transmission of data over noisy communication channels. These include: experimental measurements of the signal and noise characteristics of the magnetic media and heads, creating mathematical models from the experimental data, design of modulation, coding, signal processing, and timing recovery systems as well as theoretical studies to predict the performance of such systems.