Click for next page ( 219


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 218
7 Conclusions and Recommendations OVERVIEW The Committee to Study International Developments in Com- puter Science and Technology found that CoCom countries are maintaining a substantial qualitative and quantitative lead over the CMEA countries in computer science and technology. The range of computer technologies involved makes any single measure of that lead meaningless, but in many areas the CoCom lead is on the order of five to ten years or more. Leadership across all or most fronts is important because almost any computer technology can be use- fuT to the military, although only some computer technologies have compelling military importance. Because the committee focused its inquiry on computer technolo- gies and not on the control process, per se, it based its conclusions on the assumption that exports of computer technology would continue to be controlled. The committee's conclusions and recommendations derive from the implications of technical trends for the structure and enforceability of the control effort. While analyses conducted by the defense and intelligence communities focus on CMEA demand for computer technologies to improve CMEA military systems, the com- mittee's analysis provides complementary insights into the supply of computer technologies in CoCom, CMEA, and other countries. The committee conclu(let1 overall that technical trends make 218

OCR for page 218
CONCLUSIONS AND RECOMMENDATIONS 219 computer technology increasingly difficult to control, and this in turn increases the risk of unintended side effects of controls on the U.S. computer industry's vitality. The situation makes it important that the Department of State and other parties involved in the ex- port control process develop and administer a control regime that is more focused and more flexible than it is at present. To accomplish that, federal agencies must have more input from experts on global technical and market developments in computers on an ongoing basis. DIFFICULTIES IN C ONTRO[[ING TECHNICAL TREND S As Chapters 2 through 5 made abundantly clear, computer tech- nologies continue to be developed rapidly, and this is the most salient overarching technical trend. An important consequence is that com- puter technologies advance much faster than most bureaucratic pro- cesses. The latest round of revisions to the CoCom regulations on computer exports, for example, took several months of preparation within the United States, several months to negotiate with other countries participating in CoCom, and additional time to translate into the U.S. version of the regulations. Yet performance for many classes of affected computer systems continued to--expand during the same period. The committee's technical analyses suggest a mismatch between the rates of change (see Figures 7.1 and 7.2 and Table 7.1~. Some discrepancy is to be expected because there is a thirst rate of change to consider, that is, the change in the computer technologies of CMEA countries. Since CMEA computer technology lags behind the rate of advance of computer technology in CoCom and other non-CMEA countries, old technology outside of CMEA may be new technology within. One aspect of rapid development, the continuing reduction in the size of powerful computers and computer-related devices, means that increasingly sophisticated computer hardware is increasingly portable, easy both to hicle and to transport. It will also become increasingly widespread among both CoCom and non-CoCom, non- CMEA countries. These phenomena have been evident for some time in the realm of chips and micro- or personal computers (PCs). What is less well appreciated is the fact that other small machines, includ- ing professional workstations and minicomputer-sized machines, wiB soon be offering performance capabilities for some applications that were previously available only in conventional large supercomput- ers (see Figure 7.3~. The availability of small, high-power machines

OCR for page 218
220 M 4.50 4.00 3.50 3.00 2 50 P ~ on S 1.50 1.00 GL OBA L TR ENDS IN COMP UTER TE CHNOL O G Y )- , ~ , Hi/ 0.00 L] U US V V V V V 1 1 ~ 1977 1978 1979 1980 198 1 1982 1983 1984 1985 1986 1987 1988 | .- Microprocessors O- Decontrolled .~ Note 9 I Note 12 l FIGURE 7.1 Microprocessor performance trend versus performance trends implied by export controls, in MIPS. The top curve represents performance growth for high- end microprocessors. The bottom group of curves shows performance growth implied by OCR for page 218
CONCL USIONS A ND RE COMMENDATIONS TABLE 7.1 Changes Over Time in Threshold Performance Levels with Respect to Federal Regulations 221 Decontrol/No License National Discretion Favorable Consideration Date PDR Examples PDR Examples PDR Examples 8/1/88 6.5 IBM PC 43 IBM PC/AT 78 IBM PC/AT Intel 8088 4.77 MHz 80286, up to 10 MHz 80286, >10 MHz Mac Plus, SE 80386, up to 16 Mhz 1/29/88 6.5 28 68000, up to 12.5 MHz 48 80286, 8 and 10 MHz 80286, 6 MHz only CDC Cyber 730 DEC POP 11/73 IBM PC clones at 8 MHz 1/1/85 12/10/80 no PDR 2 Apple IIe, c 28 48 8 POP 11/03 32 IBM 370/138 DEC LSI 11 IBM 370/148 HP 2100 CD C Cyber 171 1978 no PDR 8 no PDR 1976 no PDR no PDR no PDR NOTES: Performance is measured in PDRs (processing data rates), a theoretical measure that was introduced in the mid-1970. PDRs through 12/10/80 were floating-point based only; subsequently, they applied to a maximum of fixed or floating-point operations. The governing regulation is 15 CFR 399 of the U.S. Export Administration Regulations (U.S. Department of Commerce, 1987). Notes within the regulation describe and apply to different types of computers. These notes state that licenses for export to satisfactory users in CMEA countries are likely to be approved for computers that satisfy conditions specified in each note. The microprocessors and PCa listed are merely examples of machines affected by the regulation. The national discretion columns refer to items covered under Note 9 in the regulation and its equivalent in 1978, and the favorable consideration columns refer to items covered under Note 12 in the regulation (a note that did not exist in any form in 1978 or earlier). In 1976 computers of every performance level were embargoed for export to CMEA countries. SOURCE: Data from U.S. Department of Commerce. has already increased the overall scientific computing gap between CMEA and CoCom countries. A great deal of computational power wiD be available in a great many places as more powerful smaller computers become dispersed through the scientific and industrial communities. While there are, by current standards, perhaps 1~000 large supercomputers (including vectorized mainframes) dispersed among CoCom and other non- CMEA countries today, at least 10,000 of these new, powerful smaller machines will be in use within the next five years. In some cases, clustered, networked applications of relatively small computers will

OCR for page 218
222 1 000.00 1 00.00 M 1 MOO p S 1.00 0.10 0.01 GLOBAL TRENDS IN COMPUTER TECHNOL OGY ",C - ~~"C rev - X N l~;~il I ~,,f"~ ray - 1 CDC-6600 , ~~~~~~ ' ~ ~3 6 o /6 5 I,_ :~- f,////~//~ ,3 o 3 ////' 3 0 8 ~ 11111' .~: 6 8 o 3 o to ~ ~ my 8 ~ ~ ~ 8 o 2 8 PDP-1 1 .'8088 1 1 1 1 1 1 1 1 1966 1970 1974 1978 1982 1986 I'll Supers ""at Mainframes ~ Minis ~ >~ Micros Ilililill RlSCs FIGURE 7.3 Computer performance growth. The RISC curve shows the fastest growth possible for this technology, including improvements in implementation tech- nology. Both microprocessor-based and mirucomputers are expected to move to RISC architectures, a development that will cause their performance levels to grow at annual rates of 50 to 70 percent. yield very high performance. Even though some of these machines or systems may have military importance, their anticipated affordabltity and proliferation will vitiate control efforts. Software is intrinsically difficult to control because it is easy to copy and it is supplied in media that make it highly transportable. In many cases, scientific and technical software is more or less publicly available and widely distributed throughout the scientific community. In addition, commercial software for PCs, in particular, is available through a wide variety of retail channels. Meanwhile, software is attaining growing importance as an element of computer systems. Systems for the design and manufacture of sophisticated computer hardware require sophisticated software. The increasing accessibility of PCs and other computers to nonexperts depends on easy-to-use software, and realizing the potential of advanced architectures, espe- cially for parallel processing, calls for a(lvanced software. The recent incident involving the illegal attempt to transfer to the USSR stolen Saxpy high-performance computer technology in the form of soft- ware and related technical data illustrates both the importance of software to computer systems and its transportability. CoCom has a significant lead in the areas of systems and commercial applications software in particular, and within CoCom the combination of a rich

OCR for page 218
CONCL USIONS AND RECOMMENDATIONS 223 infrastructure and an entrepreneurial culture has contributed to a significant lead for the United States. Recommendation 1: Definitions for computer technologies on the list of controlled products and their use in administering controlprograms should be made morefJexible to accountfor tech- nology change, market developments, and variations among tech- nologies that might be colloquially labeled in the same way. Fur- ther, the Departments of State, Defense, and Commerce should review definitions or categorizations of controlled technologies and their administration in a manner that is more timely and rapid, as well as more expert. The committee recognizes that advisory committees are already in use; but it is concerned that review of control decisions is neither timely nor rapid, and evi- dence suggests that more expertise may be needed to reach the best decisions and put them into eject. A key example of the need for flexibility involves super- computers, which are subject to case-by-case export restrictions. Defining supercomputers is so controversial that recent trade legislation caned for an official definition of supercomputers for purposes of export control. At issue is the capabilitythe phys- ical representation may vary and will change relatively quickly. The committee believes that a relative approach to categorizing these machines (e.g., the n percent most powerful as measured by generally accepted benchmark tests and/or as used in solving critical classes of problems) wiD work better than any abso- lute definition or label. Although technical progress will alter decisions about which machines fall into such a category, any decontrol decisions must take into account both the advancing level of CMEA technology and, as (liscussed below, whether the technology has become a commodity. On the other hand, the committee recommends relaxing controls on trade and access among CoCom (anc! other non-CMEA) countries for obsolescing high-performance computers. Many of the developments discussed above reflect a broader, fundamental trend: the proliferation of hardware and software that can be considered commodities. While the export regulations use the term "commodities" to refer generically to goods or products, the committee chose the term to draw on its economic implications. Commodity products are available in high volume and at low cost, they may be available in multiple and substitutable forms (e.g.,

OCR for page 218
224 GLOBAL TRENDS IN COMPUTER TECHNOLOGY original models and their clones), and they tend to be small and easy to transport. Commodities are also frequently available from multiple sources, often through a retail distribution channel. It is now possible, for example, to buy sophisticated chips and the most advanced PCs at discount retail chains and sophisticated PC software by mail order, inclucling mail-order rental (see Figure 7.4~. These characteristics make commodity technologies hard to control. They also help to explain why they are fundamental to the economic health of the computer industry. The committee concluded that more and increasingly powerful computer products will become commodities and therefore will be effectively uncontrollable. Recommendation 2: The U.S. government should establish and publish a list of computer technologies that are commodities, and it should promulgate a policy that exempts such commodities from controls for trade at least among CoCom nations. A com- puter technology should be identified and treater! in export control regimes as a commodity if the technology is readily available from foreign sources outside CoCom control (a condition that currently is cause for reducing export restrictions, although this is often not done in a timely manner), or if other factors (e.g., high vol- ume, low price, small size, ready availability of substitutes) make the technology effectively uncontrollable. To implement this rec- ommendation the government must establish a mechanism to identify the point at which a technology becomes a commodity. Another trend that may vitiate control efforts is the growth in computer networking. International data communications is flour- ishing among business and research entities. In particular, it is inte- gral to the operation of multinational companies in computer-related businesses, linking research, development, and production activities dispersed among several countries. Not only are individual compa- nies and organizations involved, but even individual states, acting on behalf of local industry, are exploring data communications links between local and foreign research and commercial entities. Soft- ware is probably the technology most vulnerable to covert access via computer networks (or their equivalent, the use of a modem and a phone line), although databases that describe hardware technology (such as an TC chip database) are also vulnerable. International com- puter networks probably represent the fastest growing gap between development and decision in export control strategy.

OCR for page 218
CONCL USIONS AND RECOMMENDATIONS Ik~meofFasI, Conve~ent, CourteovsSer.,ice ELEc ~ KOnICS ~r~,s SUNNYVALE ~ FREMONT 541 L`KESeE D'. 440 I'b$SION C~. <~8)733-1770 <415)770-~YS OP~ MO - AY-F12DAY 8AM-9PM SATURDAY ~ S~AY 9AM-~~ S V/SA d~ MAS TE/>CAI?D A CCEP TED ~ _ . . . _ . ~ . _ _ ~ .. ~ , ~ ~ ~ o o MAXX ~'Ca~dYato'~ Ca7p - Pb8. $7gss _~ ~MOReX ~ Te120 TAPE 1 ANe, t4 mlo. r~ w11h pUreho" of HS T 120 4 poek at '11.S2 r ~ore cost. ~N TRIMLINE ~ 14 -~ _ TELEPHONE ~ Lost Nurr~rRed~ Muto Modell,~210 . voice Volurnc Ton. or Pulse ~ ~ ItR Control Diolino ~ l~rwr . S.mphonic HQ VHS CR WIRELESS REMOTE CABLE READY | 1 1 |1 | - 'RI; l$~}lEl ~_=J .~'ven' 14Doviim.. ~ ~ : :~. t030O ~ '0 ChaAne! cot\le R-ody ~ _ _ _ D'o'ta' luner _ A . AUtO Pow~r On 0~ ~ Y . ^,,'0 r'ewme ~u~o t~ct ~ A MANUfACtU~3 /N JAPAN8YfUNAI ~ _ _ Kr~u Inru o/ 1~/ 00 _ i ~ FUJI 5~4"~ - ~T ~SKETUS 1~'4 59! ~ Box of 10 . 1 !S?SNP~ ST22S 1 201 - H~2D D~K IC~c~8 $249 r~ST~. 152~ 128v - Acc.~' $599 r M7M-Ust- 1 ~ UBl ~LACe~ ~ $995 ~ ro~n 1 My ~I Moker I Custom Labol Makor 1 L.=t $995 1 tObom . . 1 ~ ~/ORLD OF ~OMPUIERS HIMS 10MHz 286 MONO SYSTEM 512K RAM Expandable to 1MB on Molherboord S4/." 1.2MB Floppy Disk CloCk/ColanOor ~ Mono Monitor H'Q.s MonoGrophics Printer Cord s895 ~;_ ~_ - ~ C,_ ~D - ~ _ ~1 - 12 CPS L~r Ouo~ Ob*lo 630 Compot. ~ _ SPRINT Waa "O~ocmsor ~$~8 5~ 1~, . . CASIO- SCIENTIFIC ~ SOL~ CALCULAT .~, ,- Ideal tor ~h and Cornputor ___ ~ Schnce Studbe -_____ ~ C v s MO L C 100 - - - - L Cd s S ~ 888 SP~c~bns SONY0 D;DII\arl N~ 2545- 1 5~) 1 NEG CMO' WITH ,,iREE_ CAR E~OM 1SONS OCAmTT.EnADAPTER 1 I'EG 100 $29.88 ody oru onYcor Cosle Ck SAM9JNG 27C5 12 150 lol prx~oob l?ECHARGEABLE: ~ArT(RY 1 8 8 3 7 8 8 1 2 8 8 homo or COr Ul.. REMOTE CoMPATIIlLE _ _ _ _ nen FULL S'ZE.& 27C256, 250 3.88 VHS CAMCORDER ~ 1 MEGX9 256X9 6:1 Power Zoom _01_ ~ ^^ ~ - - Solid State MOS ~ '1 ~UN.s.si~ ^ d!~~ ~mooe Sensor MODULE ~U N.S..lmm High Spee Shutter $ 7 ~ Q 4 $ ~ ~ ~ DisPlay CP2250 ~ ~ ~ ~ ,~ 1 ~u ~256K~ 150 N.S. $~88 _ 1 - 1 ~ ~ . ~ ~ITECH 12SA u...~2sL~D-w" Co~, S9. m4noa_m~ 2400 ~U $8995 cc~AL$ff~9' ~ . ~ 1 ~ Cheek'ov. 1 10,000 woro'ond ~ ~;= ~ ~I'YOu r~ ~m ~a,-.ey . ~" , $3995 l ~ ti~ 1 1 ~_~ 1 l MEG X8 1OONS SIMM ~ $398 80387~20 '$698 1$468 256-150 1 NEG CMOS $8.88 EPROM 15ONS 1NEG100 SAM9JNG $37.88 APPLE MACINTOSH 11 1 30 ~ MONO SYS1'M I dM8 RAM. 800K Roppy Dbk ~ Mau I Sorld/SCS PoM~. AWh Keyboard 1 PLI 30M8 Hard Did' I MonootaphIcs w#h Monlor 1 s4295 ~ ~ . K~GTON ':. Mea_l~o ~CON~OL~ 5~_y~._ ~--~lW=F o=,. s499a ... ,. ,4 - . | ~ - . WA - A~ PC/XT 6067 AT - 2U I 88"s/6~ 317 '148 5 MIz 122 '198 .~ ~94 FIGURE 7.4 Commodity technologies, from chips to machines. 225

OCR for page 218
226 GLOBAL TRENDS IN COMPUTER TECHNOLOGY Computer networking is also fundamental to the use of super- computers, which are typically accessed by regional and even national users over wide area networks. Although there have apparently been incidents, the committee did not find evidence of a major threat from remote access by CMEA nationals, part of what is referred to as diversion-in-place. As discussed in Chapter 2, the risks that CoCom scientists and officials will detect and examine illegal use make serious diversion-in-place of supercomputers unlikely. Never- theless, security remains an important need for supercomputers as for other sophisticated CoCom computer resources. Recommendation S: The U.S. government should formulate a policy for preventing computer networks from becoming a channel for significant covert technology transfer and to protect the com- putational resources of CoCom countries. While in many cases the necessary technology exists, putting it to use may require further study or change in existing policy. For example: ~ Transborder network access is burgeoning among indi- viduals and multinational corporations via private and public networks. How is export of "soft" technologiessoftware, al- gorithms, specifications, and reports to be controlled on such networks? The question should be resolved whether it is in the best national security interest of the United States to permit CMEA nationals access to commercial and university networks, both directly while in the United States, and through remote telecommunications. Such access can and does take place. ~ Are the existing International Traffic in Arms Regula- tions restrictions on network security products necessary, or relevant to modern commercial (nonmilitary) security needs- encryption and trusted systemsin such services as banking and network-based retailing? They may retard U.S. competitiveness as non-CoCom sources grow stronger. One major trend, the movement toward paraDel processing, is likely to have a profound effect on the computer field and on CoCom's ability to prevent the CMEA countries from creating or obtaining machines with partial supercomputer performance levels. Currently, substantial research effort in Western Europe, Japan, and the United States is focused on developing hardware and software for parallel processing. As progress is made, the range of applications that wig be handled by a parallel processor will broaden. If no breakthroughs

OCR for page 218
CONCLUSIONS AND RECOMMENDATIONS 227 occur, parallel processing may remain in limited use in CoCom coun- tries, where ongoing improvements in conventional computers will provide an alternative. However, in CMEA countries without the access to advanced conventional machines, parallel processors may be a much more vi- able alternative. ParaDel processors could be constructed from low- technology components using large numbers of processors. Although the machines might be inefficient and difficult to program, they might provide capability approaching supercomputer performance levels for certain applications. The value of those machines to both CMEA countries en c! CoCom countries wiD depend on the develop- ment and acquisition of paraJ1el-processing software. Regardless of these possibilities, the committee expects the West to continue to have a substantial lead in this area. TECHNOLOGIES AS MORE THAN PRODUCTS The foregoing discussion focused on computers, software, and other products. Also of concern is the know-how needed to develop, manufacture, and use advanced computers, an aspect that may not be well protected by export controls. This distinction is important because acquisition of discrete products provides only limited tech- nology transfer. The committee concluded that the progress of CMEA countries in computing was severely limited by a lack of essential know-how. The shortcomings of CMEA countries in computer manufacturing reflect a combination of inadequate equipment and a lack of the know-how necessary for volume production of high-quaTity products. While scientific know-how is disseminated widely through open sci- entific literature, technical know-how cannot be assimilated merely by reading publicly available materials. For example, CMEA countries are known to have developed their own ion implanters for IC manufacturing, building on local un- derstanding of the fundamental technology of particle acceleration (they also obtain these machines from such countries as Liechten- stein). While in principle any ion implanter, including the least sophisticated, can be used to make any kind of semiconductor for which implantation is part of the production process, in practice the manufacture of more sophisticated semiconductors requires a high degree of reliability and quality in implantation. The fact that top- of-the-line implanters of Western origin offer higher reliability and

OCR for page 218
228 GLOBAL TRENDS IN COMPUTER TECHNOLOGY throughput than those from C MEA countries reflects the greater know-how of Western manufacturers. Similarly, the shortcomings of CMEA countries in scientific computing reflect a combination of inadequate equipment and a relative lack of the know-how that comes from experience in applying a broad and sophisticated mix of computing technologies. The spread of know-how associated with the development of computer technologies will become harder to control with the increase in standardization, another key trend. Standardization, as discussed above, is particularly important in the area of computer networking, but it is becoming more important in other software and computer hardware areas. Standardization involves common understanding of and adherence to technical goals for functionality or designs. For standards to be effective they must be published, whereas the process of developing standards requires open dialogue among the parties designing a standard to which they wiB all adhere. CMEA countries participate in international standard-setting, and they have ready access to published international standards. It is important to note that standard-setting is inherently a political process. Consequently, the outcome of the process of set- ting international computer-related standards may affect the United States differently from other CoCom countries. For example, a major trend in computer networking is implementation of the Open Systems Interconnection (OST) model of the International Organization for Standardization (ISO). Development of OST standards was strongly influenced by the communications environment in West European countries, which is significantly different from the environment in the United States. That is one reason why European vendors have been quicker to implement the standards and have been successful in exporting their products to the United States, where interest in OST has been growing. Through its influence on the competitive advantage of vendors in the United States and other countries, stan- dardization can affect the ease with which computer technologies can be controlled (absent a fully effective multilateral control process). COMMERCIAL VITALITY ESSENTIAL FOR TECHNICAL VITALITY Technology trends Lo not take shape in a vacuum. While the absolute laws of science may influence computer designs and uses,

OCR for page 218
234 GL OBA L TR ENDS IN COMP UTER TE CHNOL O G Y Recommendation 7: The U.S. government should greatly increase its investment in the monitoring of computer technol- ogy development and associated market trends around the world. Although the intelligence community monitors developments in CMEA countries, the committee recommends that more compre- hensive attention (i.e., addressing commercial as well as mili- tary applications) be paid on an ongoing basis to developments around; the world, especially in non-CMEA, non-CoCom nations (e.g., newly industrializing countries in Asia and Latin Amer- icag. Given the rapid rate of change of computer technology, the globalization of capabilities and markets, and the need to protect technology of compelling military importance, the Department of State must have a greatly expanded resource to make sound tech- nology export decisions. Further, the Department of State should undertake periodic reviews of technology trends along the lines of this report. The rapid change in computer technology makes trend presentations perishable, and thus this type of review should be conducted at least every three years. Although the Department of Commerce monitors foreign availability and performs competitive assessments, its limited resources appear to be stretched quite thin. Moreover, at this writing it had lost funding (and was seeking funds) for one of its more active information-gathering mechanisms, a computer technology watcher based in Europe. To monitor global tech- nology trends well, the government must invest in perhaps 100 or more additional, highly skilled people who are knowledge- able not only in computer technologies and their applications but also in international market trends and foreign languages. Alternatively, such talent could be tapped through a federally contracted research center or other independent source. The nec- essary monitoring would not only facilitate a focusing of export controls, but would also benefit U.S. industry and researchers. To best use technology watchers scattered among agencies, co- ordination and planning for this effort should be provided by a lead entity with a suitably broad mission. PROSPECTS FOR CMEA The premise for policy is that the United States and its allies should maintain the computer science and technology gap between themselves and potential military adversaries. But questions as to

OCR for page 218
CONCL USI ONS A ND RE COMMENDA TI ONS 235 how wide the gap should be and just how it should be maintained are open to debate. The existing East-West gap reflects several factors, of which three in particular have long retarded the computer-related technology transfer from the West to the CMEA countries. They are: 1. Export controls. For example, controls on the sale of manu- facturing know-how and facilities have contributed to the inability of CMEA countries to produce certain microelectronic and disk-storage products in large volumes and at high levels of quality. 2. Limited opportunities as seen by Western companies. For example, restrictions on direct access to CMEA markets and bureau- cratic constraints on the conduct of business, the forms of payment, and the removal of hard-currency profits from the CMEA bloc have discouraged CoCom firms from pursuing business in CMEA countries on a Tong-term and profitable basis. 3. Self-imposed constraints. For example, extreme protection- ism for CMEA firms in the forms of national security and currency controls, travel) restrictions on CMEA citizens, and poor mechanisms for internal technology transfer. The latter extend to constraints on information flow both within and between organizations and enter- pr~ses. As a result of these and other factors discussed in Chapter 6, CMEA countries have largely pursued the transfer of computer tech- nology from CoCom countries through means that are more covert than overt and more passive (e.g., reading the open scientific liter- ature) than active (e.g., by apprenticeships). Compared to a more overt en c! active strategy, CMEA efforts have had limited effective- ness. It is not difficult to imagine possibilities for significant progress if one or more constraining factors were to be removed or greatly relaxed. How might CoCom expect the continued globalization and com- moditization of computing and developments under perestroIka to combine to affect the future of East-West technology transfer in com- puting? The committee briefly considers each of the main factors in turn. 1. Export control. As discussed above, technical developments will make export controls less effective at limiting computer tech- nology transfer. New systems architectures, sustained rates of mi- croelectronic miniaturization, and the variety and rate of emergence of new products that incorporate sophisticated technologies are just

OCR for page 218
236 GL OBA L TRENDS IN COMP UTER TECHNOL O G Y some of the developments that will weaken control efforts. The globalization and commoditization of computing that have been dis- cussed throughout this report wig compound the problem; sheer physical availability of computing products and the rapid increase in the number of capable non-U.S. (including non-CoCom) suppliers are making it easier for the Soviets and other CMEA countries to find alternative technical solutions and suppliers, and to acquire and transport products. 2. Limited opportunities as seen by Western companies. The Soviets are once again touting opportunities for Western businesses. Technology transfer is clearly a major factor motivating Soviet over- tures to form joint enterprises and other activities discussed in Chap- ter 6. The Soviets would like their Western partners to transfer a great deal of technology, to train Soviet computer people, and to have the Western partners agree to arrangements that do not cost the Soviets hard currency (but, on the contrary, show them how to export and make hard currency) or other hard goods. Presum- ably, Western companies would prefer to sell products directly into the large Soviet economy while retaining the flexibility to take their earnings in hard currency as they do elsewhere. Regardless of any discrepancies in preferences, history has shown that there are always Western companies and academics willing to try to work with the Soviets under almost any conditions, and the Soviets are not unaware of this history. Right now there is anecdotal evidence of a relatively large number of Western companies and academics engaged in computer-related ventures with the Soviets. A consequence is a record amount of free or Tow-cost computer-related technology transfer. Although much of the activity focuses on transfer of fairly Tow-level technology, it is broad-based transfer, something that has been notably absent in the past. The committee also notes that because rapid technology change can limit the market life and sales volume of a new computer product, some producers of high-price, high-end products might be motivated to achieve any sale they can get. 3. Self-imposed constraints. Much of the perestroika program is designed to improve the environment for more effective development and use of technology in general, and there is a special focus on computer technology. Two important uncertainties with regard to the current reform are the extent to which the Soviets will remove or modify their self-imposed constraints, and the extent to which any such changes will make a difference. It is now harder for CMEA

OCR for page 218
CONCLUSIONS AND RECOMMENDATIONS 237 countries to keep up because there is so much more for them to keep up with. Meanwhile, the military and military-industrial entities have been priority receivers of advanced technology in CMEA countries, and there is no evidence as of this writing of any major change in their privileged position. Also, while public attention to perestrolka reforms focuses on improvements sought for the general economy, military modernization must also be recognized as one of the driving forces. In summary, Al three principal factors are now in a period of unusually rapid modification as a result of trends in computer technology and markets on the one hand and changes in the Soviet system and in CMEA behavior toward foreign contacts on the other. Other things being equal, the result will be an increase in both the quality and the quantity of technology transfer. While this situation leads to more pressure on export controls than ever before to "hold the line," as the committee has shown there are other pressures to modify export controls to be more in keeping with the technological and international "facts of life." The two need not be inconsistent. In this environment, controls can be important in maintaining the East-West gap by raising technology acquisition costs for CMEA countries. In so doing they may moderate the diffusion of computer technologies out of the military and into other CMEA communities where they could be used to cultivate greater self-sufficiency. But controls can be counterproductive when they are insufficiently fo- cused in terms of the scope of technologies affected and when they fail to take into account the truly global dynamics of computer tech- nology development and diffusion. SPECIFIC CONCLUSIONS ON TECHNOLOGY ASSESSMENTS The following specific conclusions were developed through the committee's technology assessments. They are also presented in the corresponding chapters. Hardware Advances in hardware from semiconductor devices through ma- chines of varying sizes and levels of performance continue to be rapid, with significant improvements in performance taking place in periods

OCR for page 218
238 GL OBA L TRENDS IN COMP UTER TECHNOL O G Y from one to three years. This phenomenon accelerates the movement downwar OCR for page 218
CONCL USIONS AND RECOMMENDATIONS 239 Software is less amenable to breakthroughs than hardware, but major advances are needed in software for paraded architectures. Software development is becoming more systematized and more like hardware development with components, standardization of in- terfaces, and improvements in tools that partially automate the de- velopment process. The United States is the world leader in software. One reason appears to be the favorable environment for the small-scale, creative, and entrepreneurial efforts that most often lead to software innova- tion. The United States should do everything possible to further the export of U.S. software to friendly nations to protect U.S. commercial leadership in this arena. Software is intrinsically difficult to protect because it is easy to carry and to copy. However, large packages for large systems, in particular, may be less valuable to a thief than to a purchaser because of the lack of reliable access to vendor support, maintenance, and upgrades. In some cases the usefulness of stolen software may be limited through restrictions on the distribution of source code, which is necessary for modifying software, and through encryption. The committee divided software into three principal classes for purposes of control efforts. The first is software with a compelling and direct military importance to CMEA countries, for which exports should be tightly controlled. The second is software tools that could be used to build software of compelling military importance, for which some degree of control appears necessary. The third includes ah other software, which should be freely traded within the West. Manufacturing State-of-the-art manufacturing equipment and electronic com- puter-aided design systems are the key to manufacturing state-of- the-art integrated circuits. Major changes in IC technology cause changes to cascade throughout the manufacturing process. Because of their role as enabling technologies, high-end manufac- turing technologies are especially importer to control. These tools are the key to technology leadership. This has been proven within CoCom, where Japanese gains have been associated with manufac- turing mastery. Overall CoCom advantage depends on a concerted effort to protect enabling technology developed by CoCom countries.

OCR for page 218
240 GLOBAL TRENDS IN COMPUTER TECHNOLOGY The United States is increasingly dependent on Japan for semi conductor manufacturing equipment. . Both Japan and West European countries are leaders in certain aspects of packaging. Computer Networks Networking epitomizes thee dual-use nature of computer technol- ogy. It serves as the backbone of modern military command and control systems as well as commercial office and factory automation systems and other civilian applications. Standards, particularly international standards, increasingly drive the development of network products. In the United States, product development is affected by a split between DOD-favored standards and international, commercial standards. This split has adversely affected U.S. companies' positions in network markets. The United States is a leader in design, manufacturing, and testing of protocols because of its longer experience and its widely available computer network testbeds. But this is a perishable lead that can only be maintained by continued support from government and industry and by resolution of export licensing difficulties for such noncritical technologies as DOD protocol stacks, Tow-grade encryp- tion, and computer network commodities. Security and control of access to U.S. and international research (and commercial) networks are keys to protecting CoCom computa- tional resources, but they may be inadequate. In particular, current network usage trends suggest that transborder flows of computer and communications technologies via networks are growing; transborder network access is rampant among individuals and multinational cor- porations via private and public networks. Networks provide means to export "soft" technologiessoftware, algorithms, specifications, and reportsthat may be outside the reach of current control mech- amsms. Special trade regulations (ITAR) constrain export of modern (nonmilitary) security products encryption and trusted systems- useful in such services as banking and network-based retailing. They may retard U.S. competitiveness as Cocom and non-CoCom sources grow stronger. The CMEA Countries Shortcomings of the CMEA computer technology base begin with base technologies. Design of advanced circuitry is hampered

OCR for page 218
CONCL USIONS AND RECOMMENDATIONS 241 by inadequate computer equipment, while production of sophisti- cated microelectronics suffers additionally from quality and reliabil- ity problems. As a result, certain computers are reported to rely on Western components (chips and boards). These shortcomings pre- vent CMEA countries from implementing major advances in com- puter architecture, at least in terms of volume production. High-performance computing is an area of great weakness in CMEA countries. Soviet high-speed computers are slower and gener- aLy less capable and available than those in CoCom countries. The number of machines and R&D projects in conventional supercom- puters, minisupercomputers, and massively parallel architectures is far below that in the West. This situation reflects shortcomings in the area of computer technology development, and it results in short- comings in scientific applications of computers to research in a wide variety of areas. One of the greatest contrasts between CMEA and CoCom coun- tries in general, and between the USSR and the United States in particular, is the scarcity of personal computers. Both the limited availability and the often poor quality of personal computers combine to prevent the growth of a computer-oriented culture. The current drive to add computing to educational curricula wiB change this situ- ation over time, but it is itself hampered by the scarcity of equipment and weaknesses in the overall support environment. Software has been most vulnerable to the negative effects of Soviet/CMEA economic and political systems. With only spotty ex- ceptions, none of the CMEA countries has developed strong software industries or specific areas of worId-ciass capabilities. The CMEA pattern of copying popular Western computer ar- chitectures should enable ready assimilation of associated Western software, but the inability to provide adequate computer memory and differences in end-user needs limit the value of commercial ap- plications software. Scientific and technical applications, however, are more transferable because the functions they provide are more universal. CMEA weaknesses in systems software stem from the strategy of copying from IBM and other CoCom computer manufacturers plus a scarcity of newer, more sophisticated hardware. That situation results in a lag of five to ten years or more. CMEA countries are at about the same level as CoCom in terms of the theory of programming languages, and they are also strong in the area of database management systems, although not in their

OCR for page 218
242 GLOBAL TRENDS IN COMPUTER TECHNOL OGY implementation. However, CMEA countries have not been able to develop sophisticated programming environments. The CoCom lead is widening as development tools proliferate to a rapidly increasing number of end-users. Poor telecommunications systems undermine the development and use of computer networking in CMEA countries. On one hand, Soviet research may lead the West in certain aspects of fiber opti- cal media, such as frost resistance and radiation hardness for fiber cables as well as lasers for fiber-optic transmission. On the other hand, support for high-speed computer networking is negligible, as is the quality of available service. The combination of these physical problems with cultural and political impediments to communication means that CMEA countries lack both the opportunities to test data communications advances on a large scale and the opportunities for resource sharing and collaboration that have marked use of computer networking in CoCom countries, especially among CoCom scientists. The CMEA computing industry has been marked by Tong prod- uct cycles and marginal incremental progress. The committee has not seen evidence to suggest that, even with current initiatives under perestrolka, CMEA will (lo dramatically better at meeting widespread needs and demand anytime soon, although substantial progress might be expected, partially because there is so much room for improvement. Under perestroika, the Soviets and East Europeans are pursuing a number of programs to improve indigenous capabil- ities. However, the committee also sees and expects to continue to see an intensification in the pursuit of technology transfer from the West.

OCR for page 218
APPENDIXES 243

OCR for page 218