Russian Virtual Computer Museum. → An Essay on forming the Unified System of Electronic Computers (Part I)
An Essay on forming the Unified System of Electronic Computers (Part I)
by V.K. Levin, academician, member of the Russian Academy of Sciences
translated and edited by Alexander Nitussov
In 1965–1966 design period in development of the second generation Soviet universal computers, such as the famous transistor-diode BESM-6 ( in 1967 officially accepted by the state testing commission), VESNA (accepted in 1964), M-220, URAL, MINSK, RAZDAN and some groups of control minicomputers was almost over. The computers were already coming to serial manufacturing and practical implementation.
No additional (digital) indexes are given in this review for computer numeration. They are commonly used for defining a „genealogy“ of computers, beginning from the first generation (the electron-valve machines). However, prehistory of the I/II generations is almost not considered here.
Although end dates of the works on the II generation computers were many times postponed, and their practical implementation was only a matter of future, the interest to that new development was genuine and active.
In reality, then, one couldn't say that the most consumers were unsatisfied with the situation in solving of their practical computation problems. Generally, lots of words were said about the usefulness of computers and computations in various fields, but all that had somehow „abstract character“. In fact, not many of them (if any at all) could show a concrete issue of theirs which urgently needed application of a computer, both then and even later.
There were also other difficulties and „circumstances“ hampering the beginning of serial computer production.
At the same time development of computer-based automatic control systems started. In 1966 the USSR Central Statistical Directorate (also Central Statistical Bureau) together with radio industry issued the “Pilot Project of the State Network of Computer Centres”. However, again it contained neither concrete conclusions on computer development fields nor appropriate directives and plans on organization of their production in the country.
Naturally, the foreign experience was also studied, but in rather general way and some serious attention was paid only to (then) new computer series „System 360“ of IBM corporation. Those who were working with high-performance computers liked machine CDC-6600 – that was the prototype of Cyber and Cray series. Information of the project „Solomon“ was also fascinating; that design was realized in form of supercomputer with massive-parallel structure Illiac-IV. For many years that system surpassed with its high performance all other ones, although in fact, there was only one machine Illiac-IV ever made.
The 1966 national plan of economic development already had the statement about necessity of developing pilot project of new computer line named „Raw“, which should be done in 1966-67. Contents (or idea) of the „Raw“ was defined as „... development of a complex (series) of highly reliable information-computing machines, with the performance range from 10 thousand to 1 MIPS, with unified structural and microelectronic technological basis and compatible programming systems, for computing centres and automated information processing systems“. That definition was undoubtedly authored by (or under guidance) of M.K. Sulim (1924-2000) , then chief of the ministerial department in charge of radio and electronic enterprises, producing computers of that time. M.K. Sulim himself was also computer engineer expert and designer. A year later he introduced the term „Unified computer system“, which proved to be very popular and was quickly widely accepted. Similar names were generally popular that time, e.g. „Unified System of designer documentation“, „Unified System of instrumentation“, etc. Thus, the new name was „in trend“.
The Moscow Institute of Precision Mechanics and Computer Engineering (IPMCE) headed by acad. S.A. Lebedev (1902-74); was appointed as the project leader. Every other organisation willing to participate was invited to join as a co-producer.
IPMCE started the work with issuing a report on some half-hundred pages, in 1966, in which the preparations for the new computer line (RYAD / ROW) pilot-project development were displayed.
The report pointed out, as positive, the fact that the IBM „System 360“ series of the software compatible computers with hybrid-microcircuits (therefore named „the 3th Generation“) already appeared a year before. There was also a number of foreign publications emphasizing its „revolutionary character“.
However, the scientists of IPMCE were rather skeptical about the structural solutions of the „System 360“ and its software compatibility conception in general.
In fact, the institute was not very active in that project, and even in 1966, at some high-level official technical meetings and discussions it did not put forward any generalised, or worked-out in detail, expert opinion on the ROW computers development. At that time, leading scientist of the IPMCE, and many other similar organisations, concentrated their efforts on industrial serial production of their 2nd generation computers and their implementation in modern automation systems. Each author of some computer model considered his (2nd generation) engineering conceptions to be still promising with many unrealized possibilities for further development.
At that time scientific conception of the computer family URAL-11, -14, -16, created by B.I. Rameev was the closest one to the ideas of compatible computer line. Unfortunately, his institute in town of Pensa (600 km south-east from Moscow)) was overloaded with current works on their design and industrial production, that's why Rameev's organisation was almost not involved in the „Moscow disputes“. By the way, serial manufacturing of URAL-16 began notably later than computers VESNA or BESM-6.
The designers from Minsk (now capital of Byelorussia), on the contrary, were very active in the mentioned problems. Computer enterprise of Minsk was very powerful, however, they planned to limit their production with computers of low and medium performance, which, in their opinion, were in the biggest demand on market.
The high-performance computer VESNA, was developed by a design bureau in Moscow (since 1978 re-named into Scientific Research Institute „Kvant“) under scientific supervision by V.S. Polin and the author (V.K. Levin). The programming expert team from the Moscow academic Institute of Applied Mathematics under prof. M.R. Shura-Bura was also engaged. The computers were given to Minsk plant for serial production, but it was hampered because they were a „foreign matter“ in the enterprise's philosophy, and partly in its technological processes. Nevertheless, the administration of the Soviet computer computer production branch, the USSR State Committee on Science and Engineering, military administration which needed high-performance machines and, last but not least, Moscow scientists, all objected the „Minsk line“ of only „mass-computer“ production without connection with high-performance models development. For that reason a „Minsk own“ line of the 3rd generation machines was not realised.
Rather interesting scientific developments were performed in Erevan (Armenia), but an industrial base for their manufacturing was not available there. Later the Erevan computers were manufactured by the computer plant in Kasan (capital of Republic of Tatarstan). In spite of that, a soon after the Erevan scientists managed to complete the USSR first 3rd generation development. That was mini-computer NAIRI-3. Some structural features (it was not compatible with any other machine) limited its applications, therefore, after 1975, it was „re-incarnated“ as NAIRI-4 with structure rather similar to mini-computer PDP-11 (of DEC).
In 1965 the Kiev Institute of Cybernetics (Academy of Science Ukraine) proposed its own project of computer „Ukraine“, but the Moscow experts were rather careful and even reserved because the necessary production experience for the 1st generation machines was not available there (the 2nd generation was not yet developed in Kiev that time).
Low personal interest and activity of academician V.M. Glushkov (the institute director) in the subject was another rather negative factor; leading Ukrainian cyberneticist Viktor Glushkov was one of the prominent Soviet computer scientists, his opinion was important. In fact he was then very busy with problems of automatic control systems.
Today it looks strange that the perspectives of mini-machines development did not cause real interest. Then, they only were considered either as the military on-board computers or as control machines hidden deep inside automated technological installations. More notable official attention to them (on the level of state/national scientific policy) was paid only after 1970, as resonance to appearance of American computer PDP-11 – the future prototype for Soviet mini-computer SM-4, designed under scientific supervision of academician B.N. Naumov.
The AMCM system (Aggregated Means of Computing Machinery) produced in town of Severo-Donetsk (East Ukraine) was a specific case which lies outside the subject of the present article. Its development was started in 1966-1967 by the instrumentation producing industry. According to the official statement that was equipment for automation of industrial technological processes. Their creative efforts resulted in mini-computers M-6000 and M-7000, which were quite good machines, with the structure somehow similar to the Hewlett-Packard computers. There was even some competition between them and SM-4, however, they all were in good demand on market.
New component basis was actively formed because its designers had completed their research on transistor diode circuits of the 2nd generation and the semiconductor technologies development already was in good progress in the powerful scientific centre near Moscow.
Although with some delay, however, already in the very first ES computers (computers of the unified series) integrated (monolith) circuits were implemented. At the same time IBM computers were still produced with hybrid components. Then, the academician Kamil A. Valiev was playing leading role in the field of the Soviet electronic industry in general, and B.N. Faizulaev did the same at the Scientific Research Institute of Digital Electronic Computers (SRIDEC). Before that B.N. Faizulaev headed development of transistor-diode circuitry for computer „Vesna“.
In February 1967 M.K. Sulim's department administration offered me position of the „Row“ pilot project chief. However, the condition was to develop and present the project within the next six months. For that reason the work was entrusted to our design bureau, and the structural solutions were to be oriented at the IBM's „System 360“.
At the same time that orientation was not obligatory; thus, both M.K. Sulim himself and the State Committee on Science and Technology were more interested in cooperation with French companies. Then the president of France Charles De Gaulle announced independent national development course what created much better political and economical cooperation atmosphere than difficult trade relations with the USA. At least that was believed by the USSR government.
Unfortunately, in spite of the intensive contacts the French side eventually did not (could not?) produce any reasonable concept. So, „the things didn't go“.
At the same time, it was planned to work out the measures on organisation and supply of the project works and to receive a governmental decision on providing them legal status (as a state plan) during 1967. At the beginning of 1967 the organisational forms were naturally not clear, however, broad cooperation of computer producing enterprises on the following stages of the project realization was openly recognized by the administration of that industrial branch. That caused a great tension among the computer plants directors of that time, because each of them was producing his own computer of the second generation.
It was also possible that characteristic for the centralisation tendency („concentration of efforts“) of soviet system somehow stimulated technical orientation at creation of hierarchical (developing) computer complex resembling the „System 360“ ideas. However, all that finally turned out to be not so easy as it was supposed.
Since the middle of 1966 my own preferences regarding the „Ryad“ problem were developing towards the „System 360“, or some similar computer system as a prototype. Such structures of the same type but with different names were developed by the companies RCA (USA), ICL (Great Britain) and Siemens (Germany).
At the same time there was no own practical experience of the second generation computers implementation in the USSR. In comparison with the western development the software of our (civil) computers was also „in the beginning stage“. Thus, each structure (or „architecture“, as it was named by IBM) designed by own efforts would be rather theoretical, „abstract“ and, therefore, disputable in such conditions. On the other hand, the available experience of the second generation computers development in our country was not consolidated and did not offer any decisive solution for the problem. Those were rather the reasonable projects of the leading foreign manufacturers, to be considered as possible prototypes. The fact that the big companies of several countries accepted practically same conception (software compatible computer family) looked very convincing; there was nothing specific in our country, that could make the demands to computer usage conditions principally different (and, therefore, make implementation of the foreign experience impossible).
The „System 360“ looked as well worked out in all details and prospective enough for further development of the hard- and software means, what was later confirmed by practice. Although the „System 360“ was very innovative its basic conceptions were organically derived from the the previous computer experience. For that reason it was positively accepted by the interested scientists and designers. The same can be said about abovementioned European system and RCA.
The next question consisted in software compatibility, with the aim of usage the available programs of the foreign computers-prototypes. Some were afraid that it was almost impossible and, for that reason, there was no necessity in direct reproduction of the prototypes' structures. Many suggested to improve the foreign solutions according to designers' own ideas and individual understanding, which naturally were very different.
In January 1967 a meeting of very important commission was held under supervision of acad. A.A. Dorodnitsin. The decision made there emphasized the importance (and necessity) of the prototype structure exact reproduction to create conditions for software development continuity. That statement became the basis for working out the following pilot project. Generally speaking, since then until now there was a wide range of opinions and evaluations on the state and the ways of the programming development in our country. Although analysis of the problem is not the subject of the present narration, one should notice that the basic programming conceptions implemented in created ES Computers (unified series/ the „Raw“) were widely recognized and approved. That was connected not only with attractiveness of the broad possibilities of the IBM programming system, but also with the fact that then nobody in the USSR would make serious attempt to create original, more or less functional system software (operating systems, data bases and data input-output organisation, translators, programming automation means, etc.), which was vitally necessary for working in applied programming.
Direct copy of the IBM system software at once provided high level of programming possibilities and, therefore, broad range of practical applications, including creation of various automated control systems, etc. In other words, to do what later was named „Informatisation of Society“. For that reason, on the initial stage of the ES Computers development, numerous arguing and the conflict of opinion mainly concerned not the idea of the foreign prototypes adoption but those were doubts in reality and possibility of direct programming compatibility with the prototype (incl. software) and the reality of continuous software development.
The key persons, which made decisive contribution to convincing others in reality of designing the „System 360“ compatible Soviet own ES Computers series, were undoubtedly V.S. Starkman (Institute of Applied Mechanics) and V.S. Lapin (Design Bureau for Industrial Automation (DBIA) of SRIDEC), who earlier actively participated in designing of computer „Vesna“.
Those were them, who initiated detailed study of the „System 360“ structure. Starkman focused his attention on processor units and Lapin dealt with external interface and input-output control. Their work resulted in making clear picture of available and missing technical IBM computers documentary. The missing documents were to be obtained and adapted. It was also clear for everybody, that the further work could not be done without practical acquaintance with genuine IBM computers.
In summer of 1967 the pilot project of the „RYAD“ (ES Computers) was completed. Its technical report consisted of 7 volumes, produced by DBIA and some more from several other enterprises. All documentary was demonstrated to acad. A.A. Dorodnitsin's commission, which officially approved it. That report also became the scientific and technical basis for the subsequent governmental general directive on development of computer science and industry in the USSR, adopted at the end of the same year.
The practical development, based on the IBM informational materials, was already in progress, however, there was still no common opinion on the higher administrative levels concerning the preferable prototype. Still some supported IBM while others insisted on British or German systems. The final decision – adoption of the „System 360“ - was made only in the end of 1969. Leaving aside all technical details, personal opinions and emotional reasoning (and there was a lot) one can say that the basic reasons for it were the following.
Oppositely to European variants the „System 360“ was presented with good and complete enough sets of documents. Some books – translated manuals on IBM programming and structure descriptions – were already published and could be found in many common soviet bookshops. The IBM corporation had to produce documentation and programs for the wide network of its affiliates and users, and somehow distribute it all over the world. The IBM itself did not object development of its replica – ES Computers in our country, that only strengthened its market positions, and weakened its German and British competitors on the markets where IBM itself had no -or limited- admission. At the same time IBM access into the USSR market was always closed with the embargo imposed by the American government on political grounds (it is in force until now).
Until the end of 1969 the Soviet government analyzed possibilities of purchasing licenses from the ICL and „Siemens“ for reproduction of their computers and software. Their prices were lower than of IBM (partly because their machines were of lower level). Unfortunately the needed amount of currency was not provided in proper time, the negotiations brought no practical results and numerous political conflicts of that time made the scientific and trade relations with the West rather unstable.
In this connection it should be mentioned as unfortunate misunderstanding (though popular) the fact emotionally described in popular book by B.N. Malinovskiy, „The History of Computers in Personalities“, pp 263-273, (Kiev, 1994). With all sincere respect to its author, one, nevertheless, should notice that his emphasizing of discussions about the prototype for the ES Computers („System 360“ of IBM or „System 4“ of ICL) did not reflect the real state of things.
In reality, the mentioned (in the book) choice by itself was not principally important, because all considered systems were very similar in ideas and even general design. Prominent scientist and also very respected person B.I. Rameev told B.N. Malinovskiy about sharp conflicts regarding the choice of foreign prototypes, which took place in 1969. In fact those conflicts were limited by small circle of concerned persons and were of rather personal and subjective character. Eventually, they did not noticeably influence the project development, what is actually displayed in the present essay.
By the end of 1969 technical and theoretical orientation on the IBM design became the mainstream, besides that was almost the planned time for issuing technological documents for manufacturing testing function models of our computers; the further hesitation at the choice of prototype would inevitably hamper the planned works. Sufficient amount of materials (technical and scientific) was still available only for IBM machines „System 360“. Of course, some materials were simultaneously purchased in Britain and France. Thus the institute of acad. B.N. Naumov successfully reproduced a computer of German company Siemens. Several such machines were manufactured by a computer plant, however, that was considered as some local work.
Later technical level of those European companies themselves, in comparison with IBM, was steadily declining. It means that our decision to chose „System 360“ as prototype, proved (rather by the natural current of events) to be quite reasonable.
1. In some sources also named „Institute of Precision Mechanics and Computer Technology“ (A.N.)
Published in museum, 07.02.2016