Click for next page ( 127


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 126
6 Computing Technology in the Soviet Union and Other CMEA Countries Because computing plays many roles enhancing productivity, accelerating scientific and technical progress, and even providing entertainment there is no other technological area in which the gap between Eastern Europe and the West is so pronounced. This chapter provides an overview of the Soviet computing industry, a comprehensive review of the status of major computing technologies in the Soviet Union, and an assessment of the impact of Gorbachev's reforms on the future of Soviet computing. The scope of the chapter is limited in several important ways. First, it does not try to pin- point the state of the art of Soviet military computing. The chapter is based solely on unclassified sources of information that cannot support such a review. However, the chapter does show the pa- rameters within which the Soviet military largely works and reveals that it has far less to work with than its counterpart in the United States. In addition, the limited ability of the Soviets to deploy com- puters throughout their economy influences the general ability of the economy to support the military. Second, the discussion covers mainly the Soviet Union. In some areas, the East European countries are doing better than the Soviet Union, but in general the Soviet state of the art reflects the state of the art in Eastern Europe as a whole. Furthermore, the committee could not investigate Al countries that might be candidates for export control regulations, including the People's Republic of China; but 126

OCR for page 126
COMP UTING TE CHNOL O G Y IN CMEA CO UNTR IES 127 such information about other countries has been included as could be provided, within the committee's limitations. Third, the emphasis here is more on what the Soviets have than on their potential breakthrough technologies. This approach is necessary to establish a baseline for understanding Soviet computer technologies across the board, and because specific information about potential Soviet developments and breakthrough technologies is dif- ficult to obtain. Potential breakthroughs are covered more generally in the final section of the chapter, which provides an extensive dis- cussion of the reforms being initiated by the Gorbachev regime. The success of the reforms will greatly affect the ability of the Soviet computer industry to make substantial progress. The implications of the reforms for export control are also considered. THE SOVIET COMPUTING INDUSTRY The basic division of labor in the Soviet computing industry was established in the early years of the 1960s. The principal high- level organizations in the industry are listed in Table 6.1, where their names and acronyms are given along with a synopsis of their functions. Unless otherwise noted, the organizations listed in the table were formed before or at the time of the 1965 reorganization of the economy into a centralized ministry system. With the exception of the State Committee on Computing and Tnformatics (GKVTI), which took over a number of high-level functions in 1986, and the Academy of Science's Department on Informatics, Computing, and Automation (OlVTA), which was formed three years earlier, the division of labor has remained! fairly stable (Gusev, 1983~. However, there has been a blurring of the boundaries as the three major computing ministries bring out products that overlap (Kovalenko, 1987b). Like the rest of the Soviet economy, the Soviet computer industry has been run from above on the basis of plan targets formulated by the State Planning Committee (Gosplan). In the computer industry the position of the ministries in the planning process is stronger than in many other industries; some planning functions normally under Gosplan were transferred to GKVTI in 1986. However, GKVTT did not receive corresponding financial levers over the ministries, and so during the period from 1986 to 1988 there was even less centralized coordination of the main computer ministries than before (Tvakhnov, 1988~.

OCR for page 126
128 GLOBAL TRENDS IN COMPUTER TECHNOLOGY TABLE 6.1 Principal High-Level Organizations in the Soviet Computing Industry Level Organization Functions Highest Military-Industrial Commission (VPK) State Ministry Academy of Sciences Planning Committee (G OSPLAN) Material-Technical Supply (GOSSNAB) Computing and Informatice (GKVTI) Science and Technology (GKNT) Oversees entire military-industrial complex, including ministries designated herein by footnote Until 1986, carried out all centralized planning with respect to computing Allocates most resources according to plans developed by GOSPLAN Formed in 1986, oversees most of computing industry but in a weaker form than GOSPLAN did Follows foreign developments in computing, makes recommendations for R&D programs Develops and approves standards Standards (GOSSTANDART) Radio Industry (Minradioprom) a Instrument Building, Means of Automation and Control Systems (Minpribor) Electronics Industry (Minelektronprom~a Means of Communication (Minpromevyazi) a Electrical Equipment Industry (Minelek- trotekhprom) Department on Formed in 1984; highest Academy of Informatice Computing, Sciences body that deals with and Automation computing; oversees considerable (OIVTA) amount of basic research Produces mainframes, some micros, and systems software Produces minicomputers, control systems, systems and applications software, and some microcomputers Produces chips and other components, and some microcomputers Formed in 1974; produces devices for telecommunications infrastructure Makes motors, some microprocessors, and some telecommunications equipment aMilitary-Industrial Commission oversees operations. The Soviet hardware production is divided among three main ministries, with other ministries playing supporting roles: The Ministry of the Radio Industry (Minradioprom) han- dles the production of the Unified System (ES) mainframes and has recently been branching out to build IBM-PC-compatible personal computers. It will also build one of the two standard microcomputers for the school computer literacy drive.

OCR for page 126
COMPUTING TECHNOL OGY IN CMEA COUNTRIES 129 ~ The Ministry of Instrument Building, the Means of Automa- tion, and Control Systems (Minpribor) builds minicomputers of the Small System (SM) designation. It also builds microcomputers and is in charge of a whole series of office machines that range from calculators to accounting machines. ~ The Ministry of the Electronics Industry (Minelektronprom) produces microelectronic components and several lines of micropro- cessors and microcomputers, most of which have the brand name Elektronika. ~ A number of other ministries produce essential products such as punch cards, air conditioners, and coatings for magnetic media. As many as 28 ministries are involved in computer production or supply of materials and components. Table 6.2 presents a representative list of the major hardware production facilities in the USSR. It shows graphically that infor- mation about plants in the military-industrial complex is harder to obtain thaninformation about Minpribor plants. Thelist is certainly incomplete and does not include most production facilities for pe- ripherals and auxiliary products such as floppy diskettes. There are at least three plants that produce floppy diskettes, and Minelektron- prom is said to be refurbishing seven microelectronics plants. In the early days of Soviet computing, the USSR Academy of Sciences played a prominent role in hardware development. During the late 1960s the industrial ministries took the lead in computer development as the functional duplication of Western architectures took precedence over all other programs. Most hardware-related re- search and development is now performed within institutes attached to the respective computing ministries. The USSR Academy of Sci- ences lacks equipment and experimental facilities to do more than a limited amount of hardware research. Higher educational institutions in general do far less basic or applied research than their Western counterparts. As in the Western industrialized nations, the Soviet software in- ctustry is large and spread out over a great diversity of organizations. Most ministries have designated at least one "lead" institute that oversees computerization in its branch and works with Minpribor. Some branches, such as the Ministry of the Chemicals Industry, have a large number of software development institutes scattered across the country.

OCR for page 126
130 GLOBAL TRENDS IN COMPUTER TECHNOLOGY TABLE 6.2 A Sample of Major Hardware Producers in the Soviet Union Plant Subordination Elka Plant Leningrad Svetlana Association Pavlovskiy Posad Plant "Eksiton" Voronezh Elektronika Association Zelenograd Complex Kiev Elektronmash Production Association imeni V. I. Lenin Kiev Plant of Computers and Electronic Control Machines (VUM) Kishinev Calculating Machine Plant Kursk Calculating Machines Plant Leningrad Electrical Machines Plant Livny Experimental Factory of Computer Graphics Moscow Elektronmash Scientific Production Association Orel' Computer Machines Plant Ryazan' Order of Lenin Factory of Calculating- Analytical Machinery Severodonetsk Scientific Production Association Impuls' imeni XXV s"ye~da KPSS Smolensk Calculating Machine Factory Taurage Calculating Machine Assemblies Plant Tbilisi Control Computer Works Vil'nyus Factory of Computers imeni Lenina Vil'nyus Sigma Association Vinnitsa Plant "Terminal" Kazan' Computer Plant Minsk Computer Technology Production Corporation Moscow Calculating Machines Plant (SAM) Moscow Radio Plant Penza Computer Works Yerevan Elektron Plant Minelektronprom Minelektronprom Minelektronprom Minelektronprom Minelektronprom Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minpribor Minradioprom Minradioprom Minradioprom Minradioprom Minradioprom Minradioprom

OCR for page 126
COMP UTING TECHNOL O G Y IN CMEA CO UNTRIES 131 Minpribor runs the Tsentroprogrammsistem software house and institutes that are part of the Soyuzsistemprom production associ- ation. These institutes are the backbone of the Soviet applications software industry. They are software specialized and generally have a stock of standard software that they use in the systems that they develop. These institutes farm out parts of systems (e.g., docu- mentation) to design bureaus, and it is not uncommon for three or four different organizations to work on one system. Inside enterprises there are departments for software development where maintenance of management information systems (M1:S) principally takes place. The USSR Academy of Sciences does a large amount of its re- search in basic software areas and plays a lesser role in the devel- opment of applications software. Related organizations include: the Department for Informatics, Computer Technology, and Automation; the Interbranch Scientific Technical Complex for Personal Comput- ers; various institutes of cybernetics; various institutes for economic management; various important scientific institutes (e.g., the Physics Institute imeni Lebedev); and major republic computer centers (e.g., Novosibirsk). These institutions develop applications for their own use and play an important part in defining the theory of how comput- ers should be applied. They do not have a main mission of developing software for customers. In the realm of networking and databases, a few major players stand out. The AD-Union Scientific Research Institute of Applied Automated Systems is in charge of links with Western networks and runs the Moscow portion of the network for the USSR Academy of Sciences. The Institute of Electronic Computer Technology in Riga develops the software for this system and has a large experimental computer network. The Scientific Research Center for Electronic Computer Technology in Moscow develops networking systems soft- ware, such as the Soviet equivalent of SNA. The All-Union Scientific Research Institute of Problems of the Organization of Management has been working on the network needed] to implement the nation- wide system for economic information (OGAS). Many of the other institutions that play a major role in networking are part of the USSR Academy of Sciences and are noteworthy because of the LANs they have built. The community of people involved in networking is less visible than it is in other computing areas.

OCR for page 126
132 GL OBA L TRENDS IN COMP UTER TE CHNOL O G Y BASE COMPUTING TECHNOLOGIES Component Technologies, Microelectronics, and Manufacturing Of all the Soviet computing technologies, the least information is available about component technologies. Some results of theoretical research are available, but it is difficult to determine what is being developed and what is in production on the basis of this literature. Two approaches to fining in the gaps are to work from the top down and from the bottom up. The first depends on high-level statements by Soviet officials that give an overall picture of the Soviet microelectronics scene, and the second pieces together occasional references to devices in a large body of literature. The most general assessment of Soviet microelectronics comes from the Soviets themselves. According to Yevgenly Velikhov (1987), the vice-president of the USSR Academy of Sciences and the chair- man of its Department for Informatics, Computing, and Automation, . . . in both logical design and "memory" the most fundamental work we lag by two generations. At present, we have a dearth of 64K memory media while abroad, they are beginning to sell megabyte media tchips] more than ten times cheaper. Velikhov has acted as a personal adviser to Gorbachev and is clearly in a position to make such statements. More information is provided by an unusually candid round-table analysis of the Soviet comput- ing industry, in which Velikhov and other high-ranking computer ministry officials, including the ministers themselves, participated. The leading firms have learned to place five ~layers] on 2 1/2-mm printed circuit boards, while we cannot place more than two in the same space. They are Serially-producing 36-layer wafers, whereas we call 20-layer wafers an achievement. (Kovalenko, 1987b) Furthermore, the number of areas in which the Soviets lag behind the Western state of the art is said to be in the dozens. Table 6.3 shows the results of these lags in the production of selected microprocessors and chip sets. The Soviets have actively pursued a strategy of making their own functional duplications of selected Western chip sets. Unlike reverse engineering, which implies that the "reverse engineer" has learned to make exact copies of the device, functional duplication is a weaker form of copying where the end product will function in the same manner as the original, even though it was not built in exactly the same way. The lag between the time

OCR for page 126
COMP UTING TECHNOL O G Y IN CMEA CO UNTRIES TABLE 6.S Selected Western and So~riet/East German Chip Comparison Western Chip Approximate Year of So~riet/East German Equivalent Approximate Year of Name Appearance Name Appearance Intel 4004 1971-1972 U 808 (GDR) 1978 Intel 8008 Intel 8080 1973-1974 K 580 1978-1979 Intel 300 1974 K 589 1978-1979 Chip used in 1975 K 581 1978-1979 LSI-11 AM 2900 series 1975 K 1804 1982-1983 Zilog ZB0 1976-1977 U 800 (GDR) 1980 Intel 8748 Zilog Z8000 1978-1979 U 8000 (GDR) 1983-1984 Intel 8086/88 1978-1979 K 1810 1983-1984 Intel 80286 1982-1983 No equivalent yet seen Intel 80386 1985-1986 No equivalent yet seen SOURCES: Stapleton, 1988; Stapleton et al., 1985. 133 that the original Western chip appeared and the time the Soviets were able to produce a functional duplication has consistently been about five years. One knowledgeable Soviet thinks that it would take at least three years to set up production of Intel 80386-like chips in the USSR, and that by the time production actually started, they would be obsolete. In general, silicon processing for integrated circuits is the pre- dominant microelectronics process. The Soviets grow their own sil- icon ingots and have developed their own equipment (with the help of East Germany) for processing the silicon and for manufacturing integrated circuits. The capabilities in silicon ICs are roughly at the 3-micron level versus the routine use of the 1.2- to 1.5-micron level in the United States. In the United States the cutting edge of research is in submicron feature sizes. In the Eastern bloc countries, experimental production at roughly the 1-micron level has begun at the East German Karl Zeiss enterprise. The bottom-up approach of putting together many references to individual devices confirms the statements of Velikhov and others.

OCR for page 126
134 GLOBAL TRENDS IN COMPUTER TECHNOLOGY TABLE 6.4 Partial Inventory of East European Memory Devices Western Chip Equivalent Technology Function Size K 565 RU 1A I 2107 B NMOS DRAM 4096 x 1 bits K 565 RU1B 8 kbits K 565 RU2 S RAM 4 kbits K 565 RU 2A I 2102 A NMOS SRAM 1024 x 1 bits K 565 RU 3A MK 4116-4 NMOS DRAM 16,384 x 1 bits K 573 RF 1 I 2708 NMOS EPROM 1024 x 8 bits K 573 RF 2 I 2716 NMOS EPROM 2048x 8 bits KM 537 RU 1 HM 6508 CMOS RAM 1024 x 1 bits K 555 VZhl K 556 RT 5 I 3604 TTL PROM 512 x 8 bits K 1801 RYe 1-00 NMOS ROM 4096 x 16 bits U 214 U 224 U 2316D NMOS PROM 2048 x 8 bits NOTE: U series are East German; others are Soviet. SOURCES: Konopel'ko and Losev, 1986; Stepano`', 1983; Sypchuk et al., 1983. Table 6.4 lists a number of East European memory devices ranging in size from about 1 kbit to 64 kbits. In 1985 the Soviets reported a 2-Mbit chip from the Donetsk PhysicaJ-Technical Institute. In late 1987 this was the largest re- portect Soviet chip, but it has not been demonstrated. Based on a lag of about three or four years before laboratory devices show up in the West, this 2-Mbit device would be comparable to Western developments at the time. However, since there is no further infor- mation about it, it may have remained in the laboratory because it is hard to believe that the Soviets would not be publicizing this chip if they were about ready to begin commercial production (Trud, 1985~. Devices with 4-Mbit capacity are now under development in the West. More recent chips that have been demonstrated have consid- erably smaller capacities. At the 1988 Leipzig Spring Fair, Karl Zeiss demonstrated a number of IC memory devices, including 64- kbit CMOS static RAMs, 4-kbit CMOS fast static RAMs, 256-kbit CMOS dynamic RAMs, 16-kbit Static RAMs, and 64-kbit dynamic RAMs. Memory chips that can store 1 Mbit are in the development stage. These devices probably reflect the state of the art for memory

OCR for page 126
COMP UTING TECHNOL O G Y IN CMEA CO UNTRIES 135 devices in Eastern Europe. It was asserted that series production of 256-kbit chips began in 1987 and that 40,000 had been produced by Spring 1988. It was also stated at the Fair that 4 million 64-kbit RAM chips had been produced in Dresden, but no independent con- firmation of these figures is available. Based on the table and other reports, it appears that the 64-kbit memory devices are the upper bound of what is practically available now. The memory chips in predominant use in Soviet computers probably remain the K 565 series. Velikhov and the round table pointed to a number of causes for the lag. First, the necessary design technology is not available. Velikhov (1987) says, "But given that we have little equipment, we are unable to design new generations of TCs." Designers have to "raise their hands in helplessness" because they cannot get aD the necessary pieces to fit together. The absence of supercomputers and powerful workstations for designers certainly plays a role here. In addition, engineers are subjected to a host of regulations, and the need to get approval at every stage of the process slows things down greatly (McHenry, 1987a). Second, there is a "critical inadequacy" of IC production equip- ment. Once again, Velikhov (1987) states: The capabilities of the organizations that design and manufacture the required technical equipment are not up to the task of the accelerated development of our microelectronics or the necessary rates of modernization. Velikhov asserts that the USSR production is only about 10 percent of the volume of IC production equipment in the West. In recent years the Soviets and East Germans have been more open about showing their semiconductor production equipment to the world. Much was displayed recently at the 1988 Leipzig Spring Fair.i The Soviets showed operational wafer transfer units, and automatic chip, ballL-bond, en cl ultrasonic bonders. An align and expose stepper was on display but not in operation. Large photographs of production lines using this equipment were shown, and Soviet engineers claimed that such equipment had been available for several years. An even bigger display was put on by Karl Zeiss (GDR) that showed assembly equipment, such as wafer prober, chip-bonder, 1 See Chapter 3 for an overview of the venous steps in the semiconductor manu- facturing process.

OCR for page 126
136 GL OBAL TRENDS IN COMP UTER TECHNOL OG Y reactive ion etcher, sputterer, computer-controDed equipment for analysis of defects, align and expose stepper, and electron-beam photolithographic equipment. Zeiss engineers stated that this was third-generation equipment and that they had shipped more than 60 such systems of ah generations to other CMEA users. They have shipped four or five electron-beam systems. They also stated that about 80 percent of the IC manufacturing equipment made by Zeiss is shipped to other CMEA countries, mainly to the USSR. Zeiss's promotional package at the Fair showed equipment cov- ering the entire range of VEST production, with the exception of equipment to produce high-quality wafer slices. Included were equip- ment for electron-beam exposure (ZBA 21) and devices for ion-beam etching and ultrasonic bonding using aluminum silicon wire (AlSi) and gold, respectively. These are machines associated with achieving precision improvements and cost reductions. The ZBA 21 has a pro- ductivity not much better than two 4-inch Si wafers per hour or two 5-inch by 5-inch glass-reticle-masks per hour. Some critical devices, such as the electron-beam inspection system, the photomask com- parator, the photomask inspection instrument, and perhaps even the ion-beam etching system, were shown hand drawn rather than pho- tographed, so it is uncertain whether these devices are operational and available. Certain open technical journals and academic papers reveal de- taiTs about Soviet research in areas such as materials processing, device and process modeling, discrete device manufacture, and sil- icon submicron lithography. Epitaxial growth of silicon is a stan- dard part of the Soviet IC production process, and in recent years there has been research on polysilicon for metal oxide semiconductor (MOS) gates. The Shutnikov Crystallography Institute continues to do original work on the growth of single-crystal epitax~al silicon on an amorphous substrate, a process it calls "diataxy." The use of lasers to stimulate and control silicon deposition, the use of ion-beam de- position of epitax~al silicon, and work on silicon nitrate also continue to be reported by Soviet researchers. A third reason for the lag cited by the Soviets is that the nec- essary ultrapure materials are not available to them (KovaTenko, 1987b): Materials are an even deeper source of problems. And again the problem is not that there are thousands of [materials] but rather that the majority of them should be ultrapure and have unique properties.

OCR for page 126
COMPUTING TECHNOLOGY IN CMEA CO UNTRIES 207 indicate that there is a formal proposal to shift the structure of industrial imports to provide a substantially larger supply of foreign computer equipment. This increase, if implemented, will come at the expense of chemical and petrochemical imports. The Soviet 12th Five-Year Plan gives particular emphasis to the expansion of industrial cooperation agreements, which establish a long-term relationship between a Soviet enterprise or ministry and its foreign partner. A study of 218 Soviet industrial cooperation agreements with Western firms found that 54.6 percent of them involved coproduction or specialization in production; 27.1 percent provided for delivery of plant or equipmentusually by the Western partner; 7.3 percent were joint ventures and another 6.9 percent other types of joint projects; and 4.1 percent were licensing agreements. The plans for perestrolka also provide for substantial expansion of such Tong-term industrial cooperation, which can take many forms: licensing of patents, copyrights, and technological know-how; long-term two-way trade between partners involving exchange of industrial materials and components and intermediate products; joint research and development; subcontracting for the provision of specialized services or components; sales of equipment or turnkey plants together with provision of technical assistance; sales of equipment or capital goods with deferred payment out of sales of the resulting products; and ~ joint ventures, or coproduction. The motivation to formalize patent rights at this time reflects a primary interest in acquiring Western licenses. Soviet efforts to expand industrial cooperation agreements are also directed toward Eastern Europe. At the extraordinary CMEA Council Session in December 1985, the members of CMEA elaborated a program to create mutaBy compatible technologies throughout the Soviet bloc and to set up a framework for political negotiation of eco- nomic areas of specialization in five areas: electronics, automation, nuclear energy, new materials, and biotechnology. A key feature of the CMEA investment program is the allocation of scarce invest- ment resources in the East European economies to fund long-run projects that are of particular interest to the Soviet Union. Joint specialization was already extensive in computers where the Soviet Union has Tong benefited from East European expertise. The Rand

OCR for page 126
208 GL OBA L TR ENDS IN COMP UTER TE CHNOL O G Y database of specialization agreements identifies S88 such treaties, including: 4 multilateral and 9 bilateral agreements for automated production systems and robots; 9 multilateral and ~ bilateral agree- ments in microelectronics; 5 multilateral and 15 bilateral agreements in computers; and ~ multilateral and 12 bilateral agreements for communications equipment. The Gorbachev initiatives propose increased reliance on indus- trial cooperation by means of joint ventures with foreign partners in the West as well. Although the initial regulations were relatively unattractive to prospective Western partners, the actual details of agreements worked out since the original regulations were published indicate that considerable flexibility is possible. In any event, the Soviet leadership is promoting joint ventures actively, and Western businesses are exploring the possibilities for joint production with real interest. Of the 36 joint venture agreements signed as of May 1987, 32 were with Western firms. In June 198S, the Soviet's No. 2 foreign trade official, {van {vanov, estimated that Western companies had invested $30 million in capital in 46 joint ventures since the start of l9S7, while the Soviet Union had committed more than twice as much. The original provisions for joint ventures published in Jan- uary 1987 provide that the Soviet side must hold 51 percent of the enterprise ownership and appoint the chairman of the board and director general of the firm, but the agreement may specify that cer- tain decisions require a consensus of the board members. The foreign partner may provide the official responsible for quality control. En- terprise profits bear a 20 percent tax, excluding income reinvested in the firm, and the enterprise is to be subject to Soviet labor and trade union regulations. It is not subject to central planning, but it must deal with a foreign trade organization to acquire materials directly from the Soviet economy. It can import for hard currency without re- striction and keep its hard-currency earnings, but all hard-currency imports must be covered by hard-currency exports. Repatriation of profit is limited to the amount of hard-currency earnings of the firm an arrangement that precludes import-substituting joint ven- tures. Coincident with the new regulations, 21 Soviet ministries and more than 70 large firms received the right to deal directly with for- eign firms, and the corresponding departments of the Foreign Trade Ministry were transferred to direct ministry control. Although the U.S. Department of Commerce has a long-standing policy of blocking all computer joint ventures with the Soviet Union

OCR for page 126
COMP STING TECHNOL O G Y IN CMEA CO UNTRIES . . ~ . 209 under U.S. export control regulations, both Western and Soviet press reports detail a proliferation of proposed joint ventures relating to computer products and computer controls. Some examples are listed below. The Dialog joint venture links six Soviet partners and a con- sulting group in Chicago, Management Partnerships International. The Soviet partners are the Ministry of the Automobile Industry, Moscow State University, the Central Economics Mathematics In- stitute, the Computer Center of the USSR Economics Achievements Exhibition, the Institute of Space Research, and the Vneshtekhnika foreign trade organization. Dialog hopes to make Soviet designed IBM-XT clones using Asian parts and software for sale in the Soviet Union. Its goal is 5,000 PCs during the first year, and the first lot was being assembled as of July 1988. The Dialog venture includes a production line for PCs at the Kamaz truck factory in Naberezhniye CheIny, and a factory for diskette production. Participants project annual production of up to 18,000 and sales of $160 million. Dialog recently received an unusually warm endorsement in the government newpaper ~zvestiya, and the affirmation that Soviet programming talent is worthy of a U.S.-USSR joint venture is clearly a source of pride. ~ Another joint venture links Elorg, the Soviet trade agency for computers, and three U.S. partners: New Software International in Attleboro, Mass., Innovation Computer in Cleveland, Wis., and Silicon Valley's California Micro Electronic Systems. The Western partners plan to ship IBM-PC-compatible kits to the Soviets from which they will assemble computers. ~ Wang Laboratories is talking with parties in the Soviet Union about a joint venture to build PCs in Moscow for sale throughout Eastern Europe, but it specifically says that Wang will not build its VS line of minicomputers there. . A recent issue of Datamation reports that 10 Western com- panies have come forward with proposals to set up joint personal computer production plants in the Soviet Union. Reuters has reported that the British engineering contractors Simon-Carves Ltd. have signed a 246 million contract to build a plant to make nonstrategic computers in Armenia (Radio Liberty, 1988~. ~ U.S. Applied Engineering Systems has struck a deal for pro- ducing computer control systems at the Nizhnekamsk chemical plant

OCR for page 126
210 GLOBAL TRENDS IN COMPUTER TECHNOLOGY in conjunction with the USSR Ministry of the Petroleum Refining and Petrochemical Industry. ~ U.S. Combustion Engineering is leading a consortium of Mc- Dermott International, the Soviet Ministry of Oil Refining and Petro- chemicals, and the Japanese firms Mitsui and Mitsubishi. On June 2, 1988, the Asian Wall Street Journal published an interview with Charles Hugel, president and CEO of Combustion Engineering, who said that site preparation for two huge petrochemical complexes in Tobolsk and Surgut was expected to begin this year. ~ There have been discussions between the Central Economics Mathematics Institute and IBM Europe to license the MS/DOS operating system. If any PC product is licensed, it will probably be MS/DOS. Dataease Tnternational has signed a letter of intent with an organization in Kiev. A joint venture wiD market the Dataease microcomputer database management system in the USSR if the agreement moves forward. ~ At least one U.S. training firm is vigorously pursuing the market for training trainers for MS/DOS, LOTUS 1-2-3, and other common products using Russian-language-speaking trainers. ~ PC World- USSR, a new publication from TDG, will give West- ern computer companies much easier access for advertising their products in the USSR. A Soviet agreement with the Harvard Business School wiB give Soviet firms assistance in drafting contracts and legal agree- ments with Western business firms. An agreement on some standard international contracting arrangments would help to Tower the costs of negotiating joint arrangements. An agreement between Harvard and the USSR Research Institute of External Economic Affairs wiB allow Harvard Professor Paul Lawrence and Igor Faminsky, director of the USSR Research Institute, to develop management guidelines for future joint ventures and to publish composite case studies for use in American and Soviet business schools. On the surface, these agreements give the impression of a vast opening up of the Soviet market and of firms racing to take advan- tage of it. However, it is likely that real developments will be more modest than the euphoric forecasts in the Western media because the interests of the Soviet leadership and of Western businesses are not identical. The Soviets hope that their opening to Western markets will bring them expanded hard-currency sales and credits and an in- fusion of modern technology and management skills, but they may be

OCR for page 126
COMP UTING TECHNOL O G Y IN CMEA CO UNTRIES 211 overestimating both the ability of their own inflexible organizations to assimilate new technologies at arm's length and also Western wiD- ingness to provide state-of-the-art technology and direct investment in circumstances that offer many uncertainties. The Soviet leader- ship may be responsive to the pressure for change that is sure to build up when important groups among the scientific and managerial elite begin to react to the evidence of Soviet backwardness and isolation. Western companies for their part want access to Soviet markets anct do not want to jeopardize already established sales operations in the West. Most Western businesses are wining to undertake small export- oriented ventures only as a first step in a process they hope wig ultimately give them access to the large Soviet market. Absent that prospect, they would not choose to locate microcomputer production for the world market in Minsk or Yerevan; it is easier to go to South Korea or Taiwan to set up a low-cost manufacturing operation. Few companies will want to enter barter agreements that force them to seB Soviet hard-currency-producing products in the West as a means of repatriating profits, and realistic prospects are greatest for ventures that can build on established Soviet exports. As the computer-related joint ventures gear up, they win not have an immediate impact on the supply of computer goods in the USSR, nor is it clear what will happen after they produce their ini- tially small batches. It is possible that the Soviets will then gear up production using Soviet components to avoid large hard-currency bills; but no matter how creative the joint venture agreements ap- pear, the Western partners must ultimately receive hard currency or its equivalent, and the Soviets only have so much hard currency and equivalent to spend. The computing joint ventures may have their biggest impact in transferring Western management techniques and production know-how rather than in Tong-term sales of Western products in the Soviet marketplace. Perestroika and Soviet Internal Constraints The new conditions of glasnost and perestroika are making it easier for the Soviets to use active technology transfer mechanisms. For example, it is now easier for technical people to travel to the West who are able to engage their Western counterparts in meaningful conversations. The PRC, Japan, South Korea, Taiwan, and other countries have made long-term investments in sending students to

OCR for page 126
212 GL OBA L TR ENDS IN COMP UTER TE CHNOL O G Y the United States for advanced training, and it is possible that the Soviets will start sending increasing numbers of students as well. U.S. universities wiD have neither the desire nor the means to police these students to ensure that they do not have access to selected technologies. If the Soviets decide to allow significant connections between Soviet and Western computer networks and if the West also agrees, it would be trivial for ah the nodes in networks such as BITNET, ARPANET, and others to become accessible to the Soviets. As noted above, there are already 3 minion regular users of electronic mail in the United States, and having access to these users would make for more active means of communication for Soviet scientists and engineers. What was formerly a fairly easy category to control extende(l, direct, professional contacts of a training or apprenticeship nature may now be more difficult or impossible to control. In retrospect this category of control may have been feasible only because the Soviets permitted it so infrequently, but now they are actively promoting joint ventures, which are an ideal vehicle for extended apprenticeship relationships. More potent passive mechanisms are also becoming available. The continuing development of on-line bibliographic databases in the USSR along with an increased access to ah Western publications that may be evolving as a part of glasnost wiD give Soviet developers better access to the Western literature. More and more databases are going on-line in the West, so that potential Soviet access to large amounts of information is increasing. Control of foreign access to such databases does not make much sense, because internal "front" operations can simply collect from within the United States and then send the information back to Moscow. As noted above, it is unlikely that efforts to control access will succeed in the Western democracies, although there may be successful regulations for keeping militarily related documents out of such databases in the first place. The Soviets have always used requests for proposals as a means of collecting technical data. Now that the improved economic climate has made agreements more possible, if not likely, there will be much greater use of this mechanism for collecting information about state- of-the-art Western commercial products. Increased use of overt mechanisms may increase or decrease use of covert mechanisms. On the one hand, if the Soviets can get what they want directly, they will not need to use less effective transfer

OCR for page 126
COMP UTING TECHNOL O G Y IN CMEA CO UNTRIES 213 mechanisms. On the other hand, increased contacts make greater covert collection possible. The more people sent, the harder it wit be to screen them and keep an eye on them, but those involved must also be wiring to be collectors. Rather than people who are trusted by the GRU or KGB, it is possible that leading scientists and engineers wiD be sent to the West and such people may reject specific collection targets. The increasing independence of the East European countries may also make it more Circuit for the Soviets to prevail upon individuals in those countries to serve as conduits for technology transfer. The preponderance of the information technologies in the West gives the Soviet new covert opportunities for obtaining information. The more information transmitted across phone lines in digital form, the more risk there is of sophisticated listening operations that can scan the traffic and pick out interesting sections on the basis of strings. Most users of electronic mail, for example, think of it as being as private as regular mail, when in fact multiple copies of messages may be automatically created. Since these messages identify the sealer and the receiver, it would be easier for the Soviets to associate intercepted information with its context than it is in traditional eavesdropping operations. CONCLUSIONS It is clear that there are few, if any, substantial areas in com- puting in which the Soviets have achieved parity or are ahead of the West. First and foremost, the Soviet computing industry has failed to provide a microelectronics base sufficient to match Western developments and provide for mass production of everything from personal computers to large mainframes. The resulting limitations affect every part of the computing industry, including the use of computer-aided design for developing new computers. The shortage of high-speed computing capabilities is particularly acute, hindering scientific work across the board. The Soviets also have far, far fewer personal computers than the United States, where the huge number of PCs is serving as a catalyst for all sorts of rapid developments. Weaknesses in the systems software of CMEA countries stem from a strategy of copying Western companies, like IBM, plus a scarcity of newer, more sophisticated hardware. Software engineering environments are weak. The applications software industry has been hampered by the absence of strong demand for its products, by the

OCR for page 126
214 GLOBAL TRENDS IN COMPUTER TECHNOL OG Y difficulties imposed on it by the centralized distribution system, and by the hardware itself. The telecommunications infrastructure is inadequate to support large-scale data transmission. Few wide area networks can be found, although a number of institutes have developed local area networks. Legal and political barriers and the lack of incentives for organiza- tions to talk to one another have also contributed to the Tow level of networking activity. The Soviet computing industry has been marked by long produc- tion cycles and highly incremental progress. There is little reason to believe that the industry wiB do markedly better at meeting demand anytime soon. Even if the perestrolka reforms are fully implemented, substantial investment and rebuilding are needed. Although many joint ventures are being investigated, it is unclear whether they wid play a major role in improving the Soviet computer industry. Although few unclassified sources are available that characterize Soviet military computing, it is useful to consider the bigger picture of export controls, perestroika, and the Soviet military. The commit- tee has seen no evidence that there is a comprehensive and separate military industry for general-purpose computing that is technolog- icaBy far ahead of the industry described in this report. Instead, the military has made use of the civilian computer industry, sup- plementing it as necesssary for specific products such as avionics systems. The Soviet military and military-industrial and other national security organizations have traditionally enjoyed considerable privi- leges and priorities in the Soviet computing community (Goodman, l9SS). Three of the four principal computing and communications ministries, as noted above, come under the control of the military- industrial commission (VPK). Many of the computer production fa- cilities in the USSR have had an office of inspectors for the military often active duty officers with technical training to do testing and ensure quality control on the spot. They can enforce higher testing standards for units to go to the military, and to preempt produced units from prospective civilian customers and redirect them to mili- tary or other national security end-users. Special benefits have been used (e.g., housing and Moscow residence permits) to recruit the best people from the Soviet university system. The military has also enjoyed priority in getting equipment ahead of the general economy. The national security organizations also have priority and priv- ilege with respect to foreign technology. For example, they control

OCR for page 126
COMP UTING TE CHNOL G G Y IN CMEA CO UNTRIES 215 the sizable covert collection process and the collectors themselves; they have positions within the overt trade organizations; and they have priorities with respect to resources such as hard currency. Also, overt and innocuous purchasers and users of Western technology are in no position to refuse "requests" from the KGB or important VPK or Academy organizations to inspect or borrow technology or products purchased from the West. The comrruttee has no reason to believe that this structure of technology-related privilege and pri- ority for Soviet national security purposes has changed significantly under Gorbachev and perestrolka. In fact, many regard military modernization as one of the driving forces behind perestrolka. Western acquisitions can still significantly aid Soviet efforts in at least three ways. First is the use of Western-made manufacturing equipment and technology, particularly for microelectronics produc- tion. Microelectronics is fundamental to progress in other technolo- gies, and it has a vast spectrum of military applications. In some ways, transfers of equipment and technology in this area may be more effective than for other computer-related technologies. Trans- fers are focused on a fairly small number of high-priority receivers, and assimilation tends to be primarily a technical matter; that is, the interface with the "general Soviet system" presents less of a problem than is the case with other areas like software, personal computing, and telecommunications. Second are the direct operational capabilities provided to the end-users. Reliability and raw computing power of Western devices far outpace their Eastern counterparts, and high-clout users in the military or KGB opt for these systems more often than one might think given the difficulties in acquiring service, spare parts, and upgrades. Third, specimens of computing systems have great value for copying and testing the (legree of compatibility of the duplicated hardware or software with the originals. The fraction of Western- made equipment in the total Soviet inventory may be tiny, but the fraction of Soviet and East European-made computer systems that were copied from U.S. designs is large and growing. These three categories can be abstracted as acquisition/diversion for direct pro- duction, acquisition/diversion for dual use, and acquisition/(liversion for reproduction. Two examples illustrate how Western acquisitions can have di- rect impact on a Soviet military program (Goodman et al., 1984~. The first, an example of acquisition/diversion for dual use, is the use

OCR for page 126
216 GLOBAL TRENDS IN COMPUTER TECHNOLOGY of Western samples as "place holders" for the design, development, and testing of a larger system. This permits them to move along on the larger system's life cycle before a Soviet-made substitute with the same functional characteristics is available. For example, they might use covertly acquired U.S. microprocessor and memory chips in the prefielding stages of the development of a military system. When comparable Soviet-made chips become available, they would be substituted for the U.S. circuits in the serially produced version. The second, an example of acquisition/diversion for reproduc- tion, is the use of a Western unit as the foundation and initial "host" for a Sovietized prototype. A complete copy is produced through the substitution of Soviet-made components and subsystem copies that are tested in situ on the operational Western unit. In this way, the original is rebuilt, piece-by-piece, until an entire made-in-the-USSR prototype exists. The process is more suitable for copying large electromechanical systems than for microelectronics. In addition, what the Soviets can gain from computer-related technology transfer is becoming increasingly valuable to them. The lag between East and West is now large enough that the acquisi- tion/diversion of certain Western technologies for direct production will help the Soviets tremendously. For example, if a few years ago the Soviets had acquired the production facilities for a 16-kbit DRAM chip production line, they could have supplied their computers with a newfoun OCR for page 126
COMPUTING TECHNOL OGY IN CMEA COUNTRIES 217 specialized information from them. Suddenly what was formerly a passive transfer becomes quite active. The value of transferred technology is also increasing because Soviet internal ability to exploit it is improving. The more the Soviets know, the more they are able to use what they can get from the West. A corollary is that the better the hardware they have, the more use they can make of software from the West. For example, a disk swapping algorithm was of little use when the only storage media available was magnetic or paper tape. Similarly the Soviets can make little use right now of programs distributed on CD-ROM disks or programs that require several megabytes of memory on a personal computer. Thus it is clear that advanced Western computing technologies are still quite valuable to the Soviets and that they wiD continue to be targets of acquisition. If perestrolka actually fails to bring about fundamental changes in the Soviet economy, collection efforts will surely increase. On the other hand, if perestrolka succeeds, the Soviet computing industry may finally find itself in a position to fully meet demand, and the Soviet military may find itself getting the same kind of support from its own computing industry that it now must get by using the Western computing industry as a surrogate. However, the full effects of perestroika wiB be a long time coming, and transformation of the Soviet computer industry cannot happen overnight. In the meantime, the acquisition of Western technology would seem to appear even more attractive to the Soviet computer industry as it tries to close the gap. The large Soviet investments in IBM, DEC, and HP architectures wiD ensure that they stay wedded to these designs in the future. This wiB not prevent them from jumping to their own, new technology, but it win mean that they have to be able to provide emulation capabilities for thousands of programs written for the IBM 360 and 370 operating systems, for the entire range of PDP-ll operating systems, and for at least one HP operating system. Demand for the acquisition/diversion of direct production equipment is likely to remain strong. Further, it is difficult to imagine that the Soviets will not continue to make their systems compatible with Western architectures so that they can continue to tap into the huge software market here. l:t is unlikely, therefore, that demand for acquiring prototypes to use in the reproduction process will be reducecI.