Search
Table of Contents
Visitors
Extensive experience with the Ural projects and leaps in foreign technology allowed Rameev to envision a whole new generation of world-class computing machines. Lebedev, Dorodnitsyn, Glushkov, and others were also thinking along the same lines because of the favorable conditions in the Soviet Union.
The Soviet government allocated significant resources to develop computing machinery. By the 1960s, there were dozens of computer plants and several large scientific-technical institutes in Moscow, Minsk, Kiev, Leningrad, Penza and Yerevan, which already had experience with second generation computers. Also, by the late 1960s Moscow boasted one of the most powerful scientific organizations in the Soviet Union –NISEVT. By now, the official political attacks on cybernetics were a thing of the past, and computerizing the economy, science, and industry were considered high priorities. Thus the Soviet government decided to develop the Unified Computer System series (ES-EVM as noted in Chapter One) using integrated circuits.
In the West, computers were initially developed in the United States, followed by Great Britain and West Germany. In America, IBM developed the first model 360 System in 1963–1964. This series had models with varying performance capabilities that were supported by a large selection of software. For small models, the DOS/360 operating system was proposed and for large models, the OS/360 operating system; the latter was designed because the DOS/360 was not powerful enough for large computers. The experience of developing these complex and extensive operating systems showed that they required even more labor -- thousands of man-years -- than creating the hardware itself.
Later, the British company ICL developed the third generation System-4 family of computers, which were simpler from a software point of view. Almost simultaneously the Siemens Company produced an analogous computer family.
The first country in Eastern Europe to develop a series of analogous computers was the German Democratic Republic, where the political establishment decided to copy one of the American IBM-360 models.
Behind the Iron Curtain, discussions about the structure and architecture of third generation computers began in the late 1960s. On January 26, 1967 Glushkov chaired a joint meeting of the Soviet Academy of Sciences Commission on Computer Technology, headed by Dorodnitsyn, and the Council of Ministers State Committee of Science and Technology, headed by Glushkov. There was only one item on the agenda: which ES computer system to select for development in the Soviet bloc nations? The decision was made to use the IBM-360 logic structure and command system as a prototype. Glushkov was the only opponent, and his dissenting opinion was that drawing on foreign experience would certainly be beneficial, but not to the extent of simply copying an entire system, which was already several years old.
At this time, Nikita Khrushchev put forth a government order to transfer several Soviet Academy of Sciences Institutes to various industrial ministries. The Institute of Precision Mechanics was transferred to the Radio Industry and belonged to the Academy of Sciences in name only, thus weakening the authority of its specialists in electronics technology, or to put it more bluntly -- completely undermining them.[7]
Glushkov and Rameev offered to pioneer a new computing project based on Soviet computer experience, while keeping an eye on foreign achievements. In October 1967, they wrote to the Radio Industry Ministry, which was in charge of the ES project:
The resolution to build a unified series of computers for applications in national economy and government is appropriate and timely. It requires a strong team of software engineers. Based on a foundation of streamlined technological design, we can sharply increase our production of computers, and integrate compatible computers into a variety of applications.
Success, which we hope to achieve as the result of developing a unified computer system, depends entirely on how we plan to resolve this question. It is impossible not to have serious objections to the decision to copy the IBM-360 model, proposed by the Commission on Computer Technology on October 26, 1967.
It is imperative to bear in mind that the IBM-360 system, built in 1963-1964, is already lagging behind the standards and current demands placed on mathematical machines.
...The proposal to copy the IBM-360 system is equivalent to manufacturing 1970s computers using the technical standards of the early 1960s. Considering the existing trends in science and technology, the architecture of the IBM-360 will be obsolete in the 1970s, and it will not be capable of meeting the current demands.
Copying foreign work excludes the possibility of utilizing our own collective experience of computer research, and in the immediate future, will hinder our ability to employ new principles. This will bring the development of computer technology in our nation to an end.
The design teams of Soviet mathematical machines have sufficient experience to build a family of computers that would satisfy future requirements.
...It would be a better decision to develop the architecture of a unified series of Soviet computers based on our own domestic experience and achievements, while still keeping an eye on foreign innovation.
The Ural designers had solid grounds for this conclusion. They had already produced a family of Ural-11, Ural-14 and Ural-16 – programmable semiconductor computers. A comparison of the architectural decisions and functional possibilities of the Urals with those of the IBM-360 and System-4 showed that the Soviet machines were competitive with their foreign counterparts. At this time, the Penza factory was just completing the development of the multiprocessor computer Ural-25, while Ural-21 -- a design based on integral microchips -- had been successfully fine-tuned and put into production.
The transformation to an integral microchip element base and further development of the structure and architecture of the Ural family would certainly have resulted in an advanced unified computer system. Although the Urals had a limited software library, this disadvantage would have disappeared once in mass-production and as the number of users gradually increased.
The proposal to develop ES computers received the full support of the Eastern European socialist countries. Moreover, all of them except East Germany opposed copying the IBM-360. After the negotiations in August 1968, the multilateral document "General Technical Principles of the Creation of ES Computers" was signed and approved by Bulgaria, Poland, Hungary, and Czechoslovakia – with the exception of East Germany – in which the following opinion was expressed:
The structure of the ES computer should be analogous to the structure of modern systems like IBM-360, Siemens-4004, and System-4. During the development process, it should be possible to change the structure in order to take advantage of the latest computer innovations with an accepted degree of compatibility of software and technical features, while employing domestically patented computer technology, and yet maintain the same development schedule for the entire project.
During further multilateral talks, all of the Eastern European socialist nations unanimously adopted an index of non-privileged instructions for the ES computer that matched the instruction lists of the IBM-360, Siemens-4004 and System-4. The issue of privileged instructions was discussed several times, but no decision was made. The East German specialists, who insisted on duplicating the IBM-360, suggested using its list of official instructions, while other delegations disagreed. A special multilateral meeting in November 1968 aimed at the selection of a logic structure for the ES computer did not reach a unanimous decision; therefore the problem was passed on to an Eastern bloc council of chief computer designers.
The development of Soviet computer technology did not exclude extensive international cooperation. On the contrary, its advocates – Lebedev, Rameev, and Sulim – understood the advantage of cooperating with Western European companies and consciously made efforts in that direction. Western European manufacturers of computer technology were eager to compete with IBM, given the Soviet Union's huge scientific and industrial potential. ICL took the first steps towards establishing collaboration with the Soviet Union in computer design and manufacturing, offering to share some of its System-4 technology.
Rameev was an active supporter and participant in these negotiations and he signed a series of bilateral agreements to cooperate with ICL. He figured that System-4 components could be produced by one or two Soviet computing factories, while the SRI and other construction companies could create an improved series of computers based on both domestic accumulated experience and the most recent foreign achievements.
Thus, there was every reason to conclude that the 1970s would bring new great success. But how did these events actually unfold? Why were the leading specialists, Lebedev, Rameev, Glushkov, Dorodnitsyn, and Sulim ignored, while their opponents ended up victorious in the selection of a prototype for an ES?
This problem was not discussed in the Soviet media, and it remains controversial even now. Archival materials and participants' recollections of the discussions of this issue make it possible to reconstruct the chain of events here.
The wishes of Soviet designers to employ foreign technology, mainly software libraries, were certainly clear: it was quite natural to be curious about the IBM-360 and System-4.
Yet to properly integrate the software, it was necessary to:
The Soviet Union's choice of the IBM-360 as a prototype did not match any of the above conditions, because IBM had no intention of cooperating with the Soviet Union during that period: America had placed an embargo on the sale of computers to our nation. Any documentation that was available in the Soviet Union for the IBM-360 software was incomplete because it did not come directly from IBM, but was obtained through industrial espionage. The purchase of genuine IBM-360 computers was possible only through "intermediaries," which caused enormous problems.
Soviet relations with ICL were much better, due to the efforts of Sulim, Y. D. Gvishiani – Deputy Chairman of the State Committee on Science and Technology in the Soviet Cabinet of Ministers, and other supporters of cooperation with European companies.
In accordance with a memorandum of April 26, 1968, initiated and signed by the head of ICL and Chairman of the State Committee on Science and Technology, negotiations continued in order to promote cooperation in the area of computer software.
ICL agreed to share detailed information about System-4's software with Soviet scientists and was willing to send their specialists for further assistance during development, production, and software support of Soviet-made third generation computers.
During the negotiations, which included Sulim, Rameev, and several others, ICL's representatives agreed to begin cooperative development of the next generation of computer technology. In order to compete with IBM, they prepared to commit significant funds for this joint venture and provide full documentation of System-4's hardware and software by September 1, 1969.
Excited by these promising possibilities, Rameev agreed to move to NISEVT in 1967 as the deputy chief engineer of the upcoming project; to him, the choice of the prototype seemed clear. However, the biased attitude towards both the "manufactured" success of the provincial Penza school and Moscow's monopolizing organizations -- in the first place NISEVT -- became apparent much later.
In April 1969, the Council of Chief Designers headed by NISEVT's director Sergei Arkadyevich Krutovskikh, decided that the ES required a logic structure and instruction system that precisely matched the IBM-360, despite the objections from Bulgaria, Poland, Hungary, and Czechoslovakia.
Krutovskikh based his decision on the fact that collaborative work had already begun between NISEVT and its main partner – East Germany, which was studying IBM-360's software and was vehemently opposed to taking any other approach. Chief of the Radio Ministry Industry, Valery Kalmykov, and President of the Soviet Academy of Sciences, Keldysh, backed them: the top leaders had fallen under the hypnotic influence of the proposal to avoid domestic software development.
This plan's proponents argued that IBM had the world's richest and most popular software library, which could not be rejected by even fourth generation computers; and if the Soviet Union copied the 360 series of machines, then we would be saving time and money. In December 1969 a meeting was held at the Radio Industry Ministry, at which Rameev took detailed notes as acting recording secretary.
The meeting was attended by Kalmykov, Keldysh, Gorshkov (a representative of the Military-Industrial Commission) [in Russian: Voenno-promishlennaya kommissia, or VPK], Savin, Kochetov (a representative of the Communist Party Central Committee), Rakovsky, (a deputy representative of Gosplan), Sulim, Lebedev, Krutovskikh, Gorshkov, (Deputy Chief of the Radio Industry Ministry Levin, Shura-Bura, Ushakov, Arefeva, Przhialkovsky, Matkin, and Dorodnitsyn. [8]
According to Rameev's notes, the discussion went as follows:
Sulim: Regarding the state of negotiations with the GDR and ICL.
The option of IBM-360: The GDR is familiar with the IBM-360, and they are successfully developing one of the models (R-40). We have reserves and a team able to start the project. Mastering the operating system of the IBM-360 will require 2200 man-years and 700 workers. We have no contacts with the IBM; problems could develop with the acquisition of an analogous machine, which will cost 4-5 million dollars. The GDR has only a portion of the required documentation.
The ICL: We will receive all of the necessary technical documentation and support to master it. We will need to perform some alterations for which ICL is offering to purchase a batch of its recently manufactured machines. We will be able to use our team of software engineers to design additional application programs.
A group of our programmers has already received on-the-job training at this company. There is a strong likelihood of developing fourth-generation EVMs in the near future. This company is being very helpful in all aspects, because it hopes to compete with IBM after strengthening their alliance with other European companies and us. In addition, Italian and French companies have agreed to participate in the creation of fourth-generation computer technology.
Przhialkovsky: With the IBM-360, we have a system of 6 thousand microinstructions and 90% of the diagrams for the technical-electronic memories; 70% of them have flowcharts; 7000 units have design documentation. To collaborate with the ICL we will have to abandon all of these preparations and start over, which will result in one to one and half years delay. It will also require a lot of money (to purchase the ICL computers). On the other hand, collaboration with the GDR, who is successfully working on the IBM-360, is preferable. If we strengthen our teams of mathematicians, the DOS can be operational by 1971. It's time to stop vacillating.
Krutovskikh: Our project is modeled after the IBM-360 systems. If we collaborate with ICL, the composition of the models will need to be different. The technical characteristics will change and require four to five months for a preliminary design. At ICL, there is no clear distinction among their high-end models. They are included in the series of small and medium-sized computers as supercomputers. This is best not done. To change the direction at this time will push back the deadline for preparing technical documentation by one and half to two years, or perhaps more. Having worked with the GDR on the IBM-360, we can receive DOS and OS and begin serial production, without having to develop them ourselves. The Germans have gone too far to able to start over in a new direction with the ICL. The British are only interested in competition and they will likely jerk us around. They will not collaborate with us on larger machines; plus, we cannot buy 150 of their machines right now.
Dorodnitsyn: The issue of mastering the IBM-360 is oversimplified here. In actuality, it is considerably more complex. Mastery of the OS will require no less than four years, and we don't know how useful that will be. We need to collaborate together with ICL to create our own DOS and OS and produce our own computers.
Lebedev: The IBM-360 system series was developed ten years ago. We will have to limit our line of machines to small and medium capacity because the architecture of the IBM-360 is not adaptable to larger models (supercomputers). The British want to compete with the Americans after the transition to fourth generation computers. Machines with higher productivity require more specialized structure. The British are laying the foundation for automated design. The software system for the "System-4" is dynamic, and we will be able to develop it with the Brits. In turn, this will facilitate hands-on training of our personal, which would me more effective if done while developing our own proprietary system (together with the British).
Shura-Bura: From the point of view of the system software, the American version is preferable. The OS will require modifications and we will need to be familiar with all programs to achieve that.
Keldysh: We need to purchase licenses and design our own machines; otherwise we will simply repeat what others have already done. In general, we would have to create larger machines ourselves.
Lebedev: Our mathematicians believe that our programmers would receive better training using British methods.
Rakovsky: We need to consider long-term effects and come up with a unified concept. Everyone has agreed that the IBM has the most up to date software, but its operating system is very cumbersome. It would be virtually impossible to master it in four to five years. Although difficult, a decision needs to be made right now. If we choose to collaborate with the ICL, there will be political fallout with East Germany; plus, over the next five years they will produce two hundred models of the R-40. And still, we should accept ICL's offer.
Krutovskikh: Every developer except Rameev is against working with ICL. Besides, the R-50 will be ready in 1971.
Kalmykov: On the plus side, if we have DOS, we will be able to use the machines as soon as we produce them. We can also obtain many programs from the Germans. But on the minus side, we do not have any of the IBM-360 machines, and we will not have any contact with the IBM. We will loose time if we decide to collaborate with the ICL, but we will have direct contact with them and there is the promise of collaborating on the development of fourth generation computers, which is a greater advantage. They plan to design it without the Americans to compete with IBM.
Keldysh: We should not collaborate with the ICL now, but must team up with them on the design of the fourth generation of computers.
Kalmykov: Collaboration with the ICL is not going to happen. Instead, we need to ask the Germans for more help.
Those active supporters of copying the IBM-360 were General Designer of the ES-EVM Krutovskikh, his first deputy Levin, Shura-Bura, and Przhialkovsky. If at Kalmykov's conference on 18 December, 1969 -- where the final design decision was made -- the project leader would have taken a stand against copying, the computer technology development in the Soviet Union would have taken a different path.
A few months later, the Radio Ministry approved the final proposal in favor of copying the IBM-360 system.
Sulim immediately resigned from his post as Deputy Minister; a desperate gesture of protest from a man who had done everything possible to establish contacts with ICL and understood the negative consequences of copying the IBM-360 only too well.
Rameev asked to be removed from his position as Deputy General Designer of the ES computer.
As mentioned earlier, Lebedev failed to reverse the decision; the refusal of his subsequent attempts only worsened his condition and accelerated his death.
The scientific basis for the solution to this important problem – what kind of machine the ES computer was supposed to be – was substituted by the administrative order to copy the IBM-360 system: the Radio Industry Ministry, the Soviet Academy of Sciences, and NISEVT's managers did not bother to take into account the opinions of leading computer scientists from the Soviet Union and other Eastern Bloc nations.
This decision resulted in negative and tragic consequences for Soviet computer science. In his 1991 study, Rameev analyzed the enormous labor and material costs. Following are some of Rameev's conclusions:
Up through January 1, 1989 the entire Soviet Union had 13,613 general-purpose computers. This stock consisted of:
|
Model |
First production year |
Quantity in stock as of January 1, 1989 |
Portion in overall stock in percentage |
Analogous prototype |
First production year of analogous prototype |
|
ES-1066, 1086 |
1984 |
43 |
0.3 |
IBM-3033 |
1980 |
|
ES-1061 |
1980 |
400 |
2.9 |
IBM-370/158 |
1973 |
|
ES-1060 |
1977 |
237 |
1.7 |
IBM-370/158 |
1973 |
|
ES-1055 |
1978 |
456 |
3.3 |
IBM-370/155 |
1971 |
|
ES-1046 |
1984 |
375 |
2.8 |
IBM-3031 |
1978 |
|
ES-1045 |
1979 |
1069 |
7.9 |
IBM-3031 |
1978 |
|
ES-1036 |
1983 |
933 |
6.9 |
IBM-370/148 |
1977 |
|
ES-1035 |
1977 |
1872 |
13.8 |
IBM-370/138 |
1976 |
|
ES-1033 |
1975 |
1405 |
10.3 |
IBM-370/145 |
1971 |
|
ES-1022 |
1974 |
3396 |
24.9 |
IBM-360/50 |
1965 |
|
Various computers built from 1965- 1970 |
1635 |
12.0 |
|||
|
Other imported computers |
1971-1978 |
1774 |
13.2 (less than 1% of each model) |
||
|
Total |
13,613 |
100 |
The choice of foreign analogs was derived by nominal productivity without taking into account additional parameters, which would have defined technical standards.
At first glance, the technical level of the computer inventory, expressed in years, appears to mean nothing. However, these figures hide huge differences in technical/economical indicators and effectiveness.
The use of obsolete computers and information systems wasted massive quantities of personnel, financial and materials resources, and overshadowed the technical and economic benefits they managed to achieve. The losses caused by work-stoppages ("down-time"), and trouble-shooting (low reliability) of computers and systems in 1989 cost the Soviet Union about 500 million rubles.[9]
Such were the consequences of the Soviet government's willful decision to copy the IBM-360. The "sovietization" of the IBM-360 system was the first step in surrendering the forefront position won by Soviet computer scientists in the 1950s and 1960s. The second step, which led to an even further retreat, was the mindless copying of subsequent American microprocessors, an initiative led by the newly founded Ministry of Electronic Industry. The culmination of this process was the replacement of computer research and development in the Soviet Union with importing large quantities of computers from abroad.
After receiving Rameev's letter of resignation from the position of Senior Designer of the ES computer, Kalmykov did not even bother to analyze the reasons why the country's leading computer designer and founder of the Penza scientific school made such a choice. Just like countless times before, the Communist administration failed to take advantage of this famous scientist's great creative potential, causing irreparable damage to the scientific-technical progress and to the society as a whole.
During the last years of his life, Rameev lived in Moscow. On the bookshelves in his apartment, he kept his reports, projects, and photos. Gradually, these items were donated to the Russian State Polytechnical Museum in Moscow. Rameev passed away on May 16, 1994.
[7]: The implications of this decision were enormous. Under the aegis of the Academy of Science, scientists had more creative freedom. At an industrial ministry, one worked only under orders, and creative initiatives were not usually supported.
[8] The description of this meeting and its interpretation are the opinions of the author and not of the editor, editorial consultant, or the translator. Without a doubt, the scientific proponents of the IBM-360 had their own good reasons for their arguments.
[9] Translator's Note: The official currency exchange rate in 1989 was about 0.6 rubles to $1.00.
Date: 07/02/2006 Views: 72
If we could make five to seven such computers, the Soviet Union would be in great shape.
Delicious
Digg
Reddit
Magnoliacom
Newsvine
Furl
Facebook
Google
Yahoo
Technorati