The Soviet RYAD Line of Computers
The Soviet attempt to clone the IBM S/360 line of computers
During the Cold War, the United States and the Soviet Union faced off across many fronts, with each side having its own version or counter to the other side’s developments in weapons and technology. This held true in computing as well as weapons, with both sides seeking to develop and make use of computers and various calculating machines starting from the end of World War 2. Although the computing revolution in the United States receives the most press and coverage, the Soviet Union also built and developed a wide variety of computers, and in spite of the fact that their computer technology level tended to lag considerably behind the United States, their history is just as interesting, and one that deserves to be better known.
Soviet computer designers worked under considerable handicaps, both in hardware as well as software, and it is worth studying for the ways in which they fell short just as much as for the ways in which they succeeded.
Today we will be looking at the development and deployment of the Unified System, or RYAD family of computers, which were also sometimes abbreviated as ES systems, from the Russian initials for Edinaya Sistema or Unified System. I’m going to refer to these computers as RYAD systems going forward however, as that was the primary term used for these computers in the west, and that is how the CIA referred to it in their intelligence reports.
Ryad systems were identified by an ES prior to the model number, the name Ryad simply means “row” or “series”, which means that when I say “Ryad systems” I am presumably saying “series series”. Ryad was the product of a Soviet led effort that also included Soviet allies all already joined together as the CEMA or Council of Mutual Economic Assistance, which can be thought of as a sort of civilian counterpart to the Warsaw Pact.
This system was initially designed to be compatible with the IBM S/360 series, however the Ryad is more than just a simple copy of the 360. A strong analogy may be drawn with the IBM PC compatible wars of the 1980s that we have already covered on this channel, as the Ryad was an early attempt at an IBM compatible clone, just of mainframes, not personal computers, and with the resources of multiple nations behind it rather than multiple companies.
The Soviets implemented the Ryad in their own way, using a combination of their own available technology, reverse engineering, and yes, outright copying and covertly obtained technology. Its story is an interesting one, and well worth telling as it exemplifies a lot of the struggles that the Soviet computer industry had to deal with, in their struggles to match the west in computer technology.
A quick word on sources. Much of the information I found on Ryad came from CIA intelligence papers that were declassified in the 2000s…albeit with some rather odd redactions. Since I accessed them directly from the CIA’s website, I’m probably on some sort of list now. I also made use of a book called The Technological Level of Soviet Industry, which was a collection of essays collated and published by Yale in 1977, specifically Chapter 8 Computer technology, which was written by Martin Cave and contained a number of helpful graphs and tables in addition to a vast amount of summarized details about the history of Soviet computing up to that point. I also found a couple academic papers from that era, most helpfully a paper called The Soviet Bloc’s Unified System of Computers, which was published in Computing Surveys in June of 1978.
These and the other sources I consulted gave me a lot of info, but also disagreed with one another in some areas. I have done my best to sort out the most likely answer in these areas. With regards to production numbers and related figures, precise details are difficult to come by and this is especially an area where the various sources tend to disagree, so take any number with a grain of salt. I do wonder what sources might be available in Russian on Ryad, not that that would help me much since I don’t think the dozen phrases in Russian that I know would be of any help.
Origin of the Ryad systems
The Ryad program originated due to previous Soviet failures to develop a family of computers that were genuinely upwards compatible. An upwardly compatible family of computers are ones where an entire range of computers from small to large can each run applications written for the next system down. Broadly speaking, basic systems can run only basic programs, intermediate systems can run basic programs as well as intermediate, and advanced systems can run basic and intermediate programs. Essentially, the barriers to running any given application are driven by hardware configurations, such as the amount of installed memory or the CPU speed, not software incompatibility or hardware architecture.
By the 1960s it was very evident to the Soviet Union that they were falling further and further behind the computing race with the west, with Soviet computer technology staying roughly ten years behind what was available in the west in general and America in particular. Soviet computer systems were a hodge-podge of incompatible systems with no real primary family of computers, and since support for these systems was rudimentary, each system tended to wind up so customized as its users struggled to maintain it, source their own parts, and write their own software that even systems from the same computer family frequently were incompatible with one another.
To contrast this, over in America IBM’s S/360 systems had radically shaken up customers’ expectation for computer systems, with the entire family being upwardly compatible. IBM had poured enormous resources on the development of the 360 series, spending billions on it and essentially betting the company on its success. And a success it was, with the 360 series firmly cementing IBM’s control of the mainframe market, control that they would not lose until mainframes were relegated to niche status decades later. This revolution in computing technology caused the Soviet Union to fall even further behind.
The strength of the IBM S/360 line, and the reason why they were such a huge leap forward, was that every system could run the same software so long as the hardware specs were met, which was a huge change from the status quo where different systems from the same vendor frequently couldn’t run the same software if they were different models. The status quo prior to the S/360 was that every time you changed your computer system, your software almost always had to be recompiled, rewritten, or at least modified, meaning that it was a huge additional cost on top of an already enormously expensive investment. This was a huge barrier to progress, especially as the cost of software development and maintenance was soaring well past the cost of hardware.
As a side note, an easy way of broadly seeing where a given system falls on the technology spectrum at this time, whether east or west, is to split computers into three generations. The first generation used vacuum tubes, the second generation used transistors, and the third generation used integrated circuits. There can be some blurring between the lines, as sometimes a machine could straddle two generations, but these categorizations give a fairly decent way of evaluating where a computer of this era falls. By the early 70s, the vast majority of computers manufactured in the United States were all using integrated circuits, while every single Soviet computer in mass production was still relying on transistors.
As a matter of fact, the USSR was still using quite a few first generation systems into the 1970s and struggled to get second generation systems into mass production, let alone third generation. As late as 1974 Ukraine’s installed base of computers was 93 percent second generation with the majority of the remaining seven percent mostly first generation systems. By contrast, American systems were almost entirely second generation by 1960 and by 1970 third generation systems made up the bulk of American computers with second generation systems rapidly being replaced.
The Soviets were still producing a decent amount of these second generation computers however, at least by the standards of the day. By 1972 the CIA estimated that the USSR was the third largest producer of computers in the world, with a total 1972 production of roughly 1300 computers, as compared to 3000 computers produced in Japan and about 20,000 in America.
The CIA also estimated the total install base of digital computers as around 8500 in the USSR and over 100,000 in the United States. Although placing third in computers, the disparities were extremely striking and the Soviets were becoming ever more painfully aware of them as the 1970s approached. And in fairness these numbers are probably at best educated guesses, that if anything overestimate the number of computers in the Soviet Union. For a contrasting view, a 1968 report stated that the Soviet Union lagged behind not just the US and Japan, but also France, Germany, and the UK when it came to the number of computers in service, and it seems unlikely that things could have changed that much in four years. However both estimates are just that, estimates and both may have been considerably off the mark in either direction.
Previous Soviet Computers
According to a 1978 article in Computing Surveys called The Soviet Bloc’s Unified System of Computers, from 1951-1970 the Soviet Union attempted to develop somewhere around 60 different computer systems, only about a third of which reached triple digit production numbers. Of course the early American computer industry also had a large amount of fragmentation, but by the 1960s it had shaken itself out into an IBM dominated industry with a second tier of manufacturers such as Burroughs, Univac, NCR, Honeywell and others rising and falling in opposition, or settling into specific niches that IBM wasn’t serving well or was uninterested in. However the fragmentation within the Soviet computer industry was probably at least an order of magnitude larger, especially in light of the massive incompatibilities that Soviet computers of the same make and model could possess. And the fragmentation in the United States was, as previously mentioned, rapidly reversing itself thanks to the overwhelming success of the S/360 line.
But in the Soviet Union in the mid 1960s, there was no unified computer system available, Soviet computers were hopelessly fragmented, crippled by poor technology including shoddy and unreliable peripherals, and were still dominated by first generation systems. As one source puts it, the Minsk-2 and -22 were the most widely used general purpose Soviet computers of their time. They were hopelessly deficient with respect to I/O devices and secondary storage consisted of very low performance, free-falling magnetic-tape drives.
By the time the RYAD systems were being designed in the late 60s, the most widely used Soviet computers were the upgraded MINSK-32s, an evolution of the Minsk-22 and a second generation machine with specs roughly equivalent to US computers of the late 1950s. The MINSK-32s could do roughly 25,000 operations per second, handle memory up to 64k, and was rarely capable of operating more than 20 hours at a stretch. By comparison, third generation machines such as the Ryad’s target; the IBM S/360 series; were in large scale production in the west, and could run for hundreds of hours without issue, utilize memory in the multiple megabytes, and perform millions of operations per second.
The Minsk-32 also used some peripherals imported from western Europe, and were largely incompatible between generations. It seems that around 2000 were produced from 1962-1975 and they remained in use for many years, even after RYAD systems became available.
The Soviets also had serious issues with producing software, a problem that was largely due to a lack of easily transferable compilers for high level languages such as COBOL or FORTRAN. Compilers did exist, but the incompatibilities between Soviet computers, even identical models from the same factory, made it difficult to port them between systems and made it even more challenging to develop applications that could run on a wide variety of systems. Thanks to this fact, most Soviet programming was done in assembler or even machine language, even though by 1970 there were at least two COBOL compilers, a few FORTRAN compilers and a couple ALGOL-60 compilers, to say nothing of Soviet specific compilers for Soviet programming languages such as ALPHA, ALGEC, and ALGAMS all available for pretty much every Soviet system.
Additionally, programming in a high level language was made even more difficult by the fact that Soviet computers suffered from inadequate primary and secondary storage, the undeveloped state of I/O peripherals, and the lack of effective mechanisms for dispensing software. There was also a strong bias on the part of Soviet programmers who favored the “efficiency” of machine or assembly language programming. It must also be noted that similar attitudes towards using high level languages also existed in the United States, and there were genuine speed advantages to programming to the “bare metal” of these early systems, whether American or Soviet. Still, programming at such a low level required a great deal more of the programmer, and made creating powerful applications harder.
Soviet computers of the 1960s also primarily operated in batch mode, and mostly had very rudimentary operating systems by comparison with American systems. For example, time-sharing did not exist in Soviet computers until the 1970s, everything up until that point was essentially single user.
Additional problems revolved around the fact that most Soviet systems were designed to meet specific scientific needs, and were optimized for working with numbers. By contrast, most American systems were general purpose computers that could handle not only numbers but also alphanumerics, such as letters and symbols. This allowed American systems to leap even further ahead in data processing, opening up new avenues for computer usage.
In fairness it must be stated that there were Soviet computers prior to RYAD that were more aimed at general purpose computing, with the Ministry of the Radio Industry formally responsible for overseeing the production of general purpose computers, although the overlapping spheres of Soviet ministries also meant that the Ministry of Instrument Construction, Means of Automation, and Control Systems as well as the Ministry of the Electronics Industry were also involved in the production of computers, including general purpose ones. There may have been other ministries involved as well, plus efforts in the various Union Republics, meaning that there was extensive fragmentation and duplication of effort.
The Soviets knew this, and knew that they needed to make a giant effort to close the performance gap. This was apparently a topic of discussion by Soviet planners as they worked on the requirements for the eighth Five Year Plan that would run from 1966-1970. Although no public announcement of this fact was made at the time, possibly due to a desire to avoid admitting that a significant computing gap existed, a formal decision was made to create an entirely new system of Soviet computers that not only covered a whole range of compatible systems and peripherals, but also were optimized for general purpose computing. The decision may have been made as part of the official planning for the eighth Five Year Plan but it was not publicly spoken of until 1967.
It seems that another impetus to develop the RYAD systems was the failure of the Ural-10 series, comprising the Ural-11 and Ural-14 in 1965, and the Ural-16 in 1969. These were updated versions of an existing line of Ural computers produced between 1956 and 1962, one of which was a young Charles Simonyi’s first computer experience in Hungary, an experience that sparked his interest in computers and eventually led him first to Xerox and then to Microsoft. That fact has nothing to do with this video, but is a neat bit of trivia.
The Ural-10 series failed to get any real traction, although serial production of the Ural 14 and 16 continued into the 1970s. According to one source, a major reason for their failure was that even by Soviet standards they were difficult to program, expensive, and poorly designed. Even though the Ural was in production in 1967, and still had one more model to be released, it was clear that this wasn’t the family of computers that the Soviets were looking for, and a new approach would be needed.
The plan was to spend three years developing this new family of computers called RYAD, and then put them into mass production in 1970. This would require the Soviets to make a significant advancement in their technology in a very short amount of time. According to a CIA analysis, integrated circuits were not being produced serially, nor were multi-layer printed circuit boards, which together with integrated circuits, are the fundamental building block of third-generation computers. The magnetic disc technology, essential for all RYAD models, was only in early stages of development. Ryad was supposed to only close the hardware and software gap with the west, but also give the Soviets a way to run the vast and rapidly expanding body of western software available for western computers.
IBM’s S/360 series was certainly not the only successful third generation series of computers in the west, as IBM’s competitors, colloquially known as the BUNCH also had third generation systems in production, some of which were superior to the 360 series in performance. However the 360 series was by far the most widely available, the most well known, the best documented, and its peripherals and various components were also manufactured by a number of third parties. This meant that it would be the easiest system to study, and one source even states that the Soviet Union was able to legally buy some 360 systems, although the details of the how are unclear and I also found references to them illicitly acquiring them as well.
In 1968 the Soviet leadership pressured all CEMA countries to join the RYAD development effort, with Bulgaria, Czechoslovakia, East Germany, Hungary, and Poland all reluctantly joining the effort. They were reluctant to increase their dependence on the Soviet Union, especially in such a key area as computing where a number of these nations were already making good progress independently, especially East Germany. Still, with Moscow putting the pressure on, they had little choice but to join the development effort and commit to purchasing systems.
Decision to base Ryad on the IBM 360 series
Interestingly, the decision to make the Ryad computer line compatible with the S/360 series was not the first time Soviet planners had tried to achieve S/360 compatibility. Several years prior to the announcement of the Ryad program there was a failed line of Soviet computers called the ASVT, which is a Russian abbreviation for Modular System of Computers. The ASVT line eventually went into production in 1968, with several models, specifically the M-2000 and the M-3000 aiming for compatibility with the S/360 instruction set. Unfortunately, Soviet circuits had not achieved the level needed for this, and the project was a total failure, at least from a compatibility standpoint. Although it overlapped with the Ryad program, it was clearly a dead end and resources were shifted away from it to Ryad.
According to a CIA overview of the Ryad program that was declassified in 2000, by modeling RYAD computers after the IBM 360 series, the USSR hopes to save both time and money and make use of the large stock of IBM software. The document goes on to say that Although the concept was sound, the program has been hampered by shortages of high-quality components, out-of-date production and testing techniques, and absence of effective direction and coordination.
The Soviet Union planned to be producing 3-5 thousand RYAD systems per year by 1975, and expected RYAD systems to be disseminated widely across not only the Soviet Union but also the Warsaw Pact countries. Unsurprisingly in light of the high ambitions and tight deadlines, everything started to slip relatively quickly as Soviet engineers began wrestling with translating the S/360’s architecture and instruction set over to something that could be achieved with the Soviet technology of that era, together with attempting to push the state of Soviet technology forward and moving technologies such as integrated circuits out of small production lots and into full mass production..
Ryad System Planning and Production
The main RYAD design and planning was done from the Scientific Research Center for Electronic Computing Technology, which was headquartered in Moscow. Basically from day one the program began running behind schedule, hindered by language barriers, struggles adapting western components and designs to available Soviet technology, massive distances between the different centers working on Ryad, poor communications, and the inevitable problems arising from trying to coordinate many different entities that had never attempted to work together before.
It wasn’t until 1971 that the first Ryad system, the low end ES-1020 was shown to the public, with production beginning in January 1972. The initial small production run apparently had issues with I/O timing, and thus were recalled within two months for corrective work.
In May of 1973 six out of the seven final Ryad systems were shown in Moscow at the Exposition of Achievements of the National Economy. Out of these six, four were more-or-less compatible with the IBM System 360 instruction set, the ES-1020, 1030, 1040, and 1050. But struggles with production were still quite clearly evident, as out of the six displayed, only three were claimed to be in actual production. It took another year before enough of the Ryad systems had gone into serial production to be considered a modest success, as by the end of 1974 Ryad systems were in serial production across the USSR and the Warsaw pact, although Ryad systems with reasonable IBM S/360 compatibility were only made in Bulgaria, Poland, East Germany, and the Soviet Union itself of course, which unsurprisingly seems to have made the bulk of the Ryad systems. Hungary and Czeckoslovakia’s Ryad systems were only partially S/360 compatible.
This chart shows the main Ryad systems, the countries that made them, and the primary characteristics and peripherals of each. Keep in mind that there is significant disagreement about how accurate these numbers are, the note at the bottom of this report says that these figures were derived from multiple presumably Soviet sources, and that there are some significant differences among the numbers given by these sources.
Even once serial production began there were some serious quality problems, the main Ryad production plant at Minsk being a prime example. According to one source, Ryad systems were assembled there with an overriding consideration for meeting production goals rather than quality, hence systems came down a conveyor line very reminiscent of an automobile production line, and all stations were allotted the same time interval for task completion and the whole line was advanced at the same time. Faulty parts or incomplete tasks were simply tagged and assembled into the product-corrections were to be made later. This meant that many Ryad systems were marked as completed while actually being non-functional and needing additional work to repair or replace faulty or improperly installed components.
The Soviet centrally planned economy also created bottlenecks in various areas of computer manufacturing and support that affected Ryad production, just as it did all computer models. Computer manufacturers would receive a priority list of components to build and provide, with hefty punishments for missing the primary quota and rewards for fulfilling it. However the plant would not want to overfill its quota as that would result in a higher quota the following year, which would be harder to meet and thus potentially result in lower pay or even punishments.
Soviet’s workers’ pay was heavily affected by what was called “premia” or bonus, which could be as much as 30 percent of their entire compensation. Hence the difference between reaching 99% and 100% of the assigned quota could mean a massive reduction in the worker’s pay, all the way up the line to management. This meant that plants walked a delicate line between filling the primary quotas so they could get the full compensation, but not overfilling them and making the following year much harder. In the area of computer manufacturing, this could mean that a plant that received a priority list of CPUs, and a secondary priority list of peripherals, would focus most of its efforts on producing the CPUs and far less on producing the peripherals. Supporting the finished product, especially if the finished computer equipment was being sent a long way away, would be far from a high priority. And customer service, if it appeared on the list at all, would generally rank near the bottom if not at the bottom.
Due to the significantly increased complexity of the Ryad systems, Soviet planners attempted to improve this situation by setting up centralized repair services for hardware, however issues persisted and replacement parts for Ryad systems could take up to a year to be delivered, as opposed to days or less in the United States. The issues got worse the further away the computers were from major metropolitan areas, such as an Uzbekistan installation of a ES-1020 Ryad system that spent years in a room with no air conditioning, which caused the computer to regularly overheat and crash, with an uptime of just three hours a day.
Ryad Peripheral Production and Design
Different members of the Warsaw Pact specialized in different peripherals for the Ryad family. For example, Poland specialized in producing line printers, Hungary focused on graphics devices, and Bulgaria became the key supplier of disk drives. A good example of this was an East German ES-1040 that Control Data Corporation managed to purchase in 1975 for testing and evaluation. CDC was interested in potentially manufacturing plug-ins for the 1040 line, and thus did a very thorough teardown and evaluation.
The 1040 was found to be a fairly decent collection of parts from around the Soviet Union and Warsaw Pact countries, including a Hungarian graphics display, disk drives that were probably from Bulgaria and a line printer that was presumably from Poland. For the record, I cannot confirm this, as the Hungarian graphics display is the only piece with a specific provenance given, but Ryad systems were supposed to be composed of parts from around the Soviet Union and allied countries so it's logical to assume that the CDC’s test machine fit that. The CDC evaluation found that the CPU’s design and reliability were good, but the RAM speed was not properly matched to the CPU’s speed of access, a common problem with Soviet systems. The 1040’s RAM also required about twice the amount of power that a comparable IBM S/360 would have required, and was prone to overheating, hence it was run at below its designed speed. However the 1040’s I/O technology was found to be fast enough and compatible enough to work with most IBM S/360 peripherals.
As Ryad systems began to be delivered, peripheral production did remain a sore spot, in 1973 only about half of the peripherals announced for Ryad even being available for order, let alone delivery with a system. Hungary was apparently the most successful of the CEMA countries in supplying peripherals, while the Soviet Union was surprisingly the least successful.
Western Technology in Ryad
Western technology played varying roles across the CEMA countries, with Soviet satellite countries making far more direct use of it in their Ryad designs than the Soviet Union itself did. The satellite CEMA countries that were designing Ryad systems instead of just peripherals or contributing to overall research were only responsible for designing a single system, and some had greater access to western components in bulk, making their part of Ryad a bit easier although still very difficult and technically challenging.
It appears that prior to the start of the Ryad program, the Soviet Union got a hold of an unknown number of S/360 systems, and East Germany itself clandestinely got a single S/360 on its own. These systems provided the Ryad program with significant research and reverse engineering capabilities that helped tremendously.
Nevertheless the two main technical countries developing Ryad systems, the Soviet Union and East Germany, used their S/360 systems in differing ways. The Soviet Union chose to focus on copying the IBM architecture, while East Germany decided to copy the physical layout, subassemblies, and components. Copying by East Germany, which involved dismantling, measuring, and analysis of parts and circuitry, required at least 7,000 scientific-engineering man-years to bring the ES-1040 program to the stage of production prototypes.
This difference in approach was mostly likely due to East Germany’s greater or easier access to western components, while the Soviet Union figured that designing systems that implemented IBM’s architecture entirely using Soviet components would save valuable time.
The single Ryad system that the East Germans designed, the ES-1040, was the most S/360 compatible Ryad system available, but did make use of at least some western components and was the result of far more duplication than reverse engineering. According to the CIA, East Germany also received a lot of assistance from West Germany, including equipment and technical knowledge, things that helped their Ryad development immensely.
According to another declassified CIA analysis of the Ryad program, Western technology acquired both overtly and covertly by the Communist countries over the past 15 to 20 years as part of their overall electronics industry buildup clearly has been a major asset in their ability to build the Ryad computers. Nevertheless, the selection of components and circuits for the Ryad computers to date was influenced more by local experience and technical supplies than by Western technology acquired specifically for Ryad projects.
It seems that the Ryad systems manufactured in the Soviet Union itself relied on Soviet manufactured components, which were definitely influenced by western technology both overtly and covertly acquired, but were engineered using available Soviet technology. Soviet Ryad systems were built using Soviet components and did not rely on materials or components imported from the west, although this ignores the impact of the Soviet Union getting its hands on critical western machinery for making computer components, more on that in a minute.
Rather entertainingly, Hungary’s Ryad system, the ES-1010, apparently was a licensed copy of a French computer that Hungary had, in 1970, gotten a full legal license to produce, together with all the technical data needed. Hungary simply modified this system to accept Ryad peripherals and as near as I can tell, just ignored S/360 compatibility entirely.
Western machinery was also acquired to help manufacture components, and a large increase in Soviet memory production and integrated circuits volumes seems to have resulted from this, with the Soviet Union getting its hands on entire plants for making circuit boards, and machinery for testing memory among others. One CIA analysis states the illegal acquisition of core presses and test machinery made it possible for the USSR to provide increased memory capability in its new model RYADs developed after 1974. So while the Soviet Union built Ryad systems using domestically available technology, it was technology that was in some cases built by machines directly sourced from the west. Production and use of Ryad systems continued to face significant bottlenecks in both hardware and software.
Software and Hardware bottlenecks
By the 1970s, the software industry in the United States was mature, and comprised everything from large companies all the way down to single programmers working on a pet application to fill a very specific niche. The only significant difference between it and the software industry of the following decade was that the 1970s software industry was still primarily focused on mainframes and minicomputers. The Soviet Union on the other hand struggled to build up a software industry and never attained the same level of success as in America. As a matter of fact, up until the Ryad ecosystem, Soviet computers were pretty much delivered and set up, and then the business was left to figure out how to not only make use of the machine by writing programs for it from scratch, but also had to acquire their own spare parts and keep engineers handy to repair any issues as little to no support from the manufacturer would be forthcoming.
Additional bottlenecks occurred due to the Soviet Union’s lack of automated machinery for producing computer components. According to one source, Lacking the advanced machinery of Western producers, the Soviets have substituted manual techniques for automatic wiring of back panels and for automatic insertion of components into printed-circuit boards. In addition, the Soviets have relied on simple electrical and electronic instrumentation (such as voltmeters and oscilloscopes) for online testing of subassemblies, while most Western producers use computerized test equipment.
There was also a severe lack early on of available integrated circuits as the Soviets struggled to produce them in anything but small batches, most of which went to the military. However this ceased to be an issue after 1973 as additional production was brought online and yields greatly increased.
Software continued to be a major problem for the RYAD throughout its life, as although the available RYAD operating systems were centrally stored and distributed, continued improvement and development of them was spotty. Additionally, software libraries of useful applications were extremely lacking, with attempts to set them up marred by a host of issues including a lack of proper staffing, poor quality control, and limited indexing that made it difficult for users to find applications in the first place, meaning that the benefits of S/360 compatibility were greatly reduced. Still, it was an improvement over the previous status quo.
The Ryad systems were also supposed to enable large scale time sharing and remote access, everything ran well behind schedule. The plan had been to have twenty Ryad centers set up for timesharing by 1975, however by 1977 not one of these centers was fully operational. The poor state of Soviet telecommunications also played a large role in this, it was not entirely a computer problem. The Soviet telephone system was incapable of supporting large-scale remote data-processing operations even if US-quality terminals, modems, etc., were available in unlimited quantities. As a matter of fact, what Soviet telecommunications did exist in the 1970s apparently relied heavily on telegraph lines, with a usable bandwidth rate of a mere 50-100 bits a second.
Another significant bottleneck that Ryad systems faced as they began to be delivered was disk storage. 1973 Ryad systems could get 7.25 megabyte removable disk packs, but that was the maximum size available. Four years later, largely thanks to the efforts of the Bulgarians, 30 megabyte drives were becoming available, but reliability remained problematic. The Bulgarians focused their efforts on driving Soviet disk technology forward, and had the most success of any CEMA country in the 1970s, even producing prototype 100 megabyte disks by 1978, although mass production of these disks still was far from being solved by that time.
It should be noted however, that Bulgaria seems to have gotten significant assistance from West Germany in advancing its disk drive technology.
Disk peripherals seemed to be the area where the Soviet Union and its allies sought to import the most peripherals from the west or at least get technological help. Wherever they came from, the Ryad systems benefited from the best peripheral ecosystem in the Soviet Union, and were the first Soviet large scale systems to take at least some steps towards ease of use. Ryad card readers were slow by western standards, but prior to Ryad the standard was paper tape or via card readers that were so sensitive to thickness and humidity that they would crush decks and jam so often as to be effectively out of commission half the time.
Other Ryad peripherals that were a dramatic improvement were tape drives, better punch card stock, and line printers that could handle cyrillic characters.
RYAD maturity and usage
According to one source, with the maturity of the Ryad family of computers, the Soviet Union could be considered “the second-largest computer company in the world. The Soviet Union alone, and the CEMA countries together, have assets that no other computer organization except I.B.M. has: huge capital and personnel resources and centralized management control involved on a large scale with the full spectrum of computer-related R&D products, and services.” This quote also points to one of the real weaknesses of Soviet computers, in that it took the entire Soviet Union and CEMA countries just to be considered in the same breath as IBM.
Ryad systems could run a total of four operating systems, although it does not appear that all of them could run on any machine, rather the system specs of a given installations’ Ryad computer would determine their operating system options. The four systems were OS 10/ES, intended for use with the lower end Hungarian made ES-1010 and only capable of running a single program at a time, MOS/ES aimed at the Czech ES-1021, DOS/ES which was a disc based OS for ES-1020, 1030, 1040, and 1050 systems with memory capacities between 64k to 128k, and finally OS/ES which ran on the same systems as DOS/ES, but required quite a bit more installed memory. OS/ES systems required a minimum memory installation of 128k. In order to run S/360 programs, it appears that the Ryad system had to be running either DOS/ES or OS/ES.
Interestingly, the two main Ryad operating systems were not homegrown Soviet ones, and hence their ability to run S/360 programs was not an original creation but rather a benefit of the hardware compatibility Ryad had aimed for. The reason DOS/ES could run most S/360 applications was apparently because DOS/ES was actually functionally identical to the IBM S/360 DOS operating system, specifically S/360 Release 26 from December 1971.
Likewise, OS/ES was a copy of OS/360, specifically IBM Release 21 from August 1972. These two operating systems were not re-engineered, reverse-engineered, or reworked IBM systems, they appear to be actual copies, with one source saying that assorted error codes, messages console commands, and software diagnostics were originally in English and identical to those used by IBM. These expressions have since become available in Cyrillic.
Oh and the low end OS 10/ES that was aimed at the lowest specced Ryad models wasn’t original either, it was actually a port of the French Mitra-15 operating system. So out of the four operating systems, as near as I can tell only MOS/ES was actually original to the Soviet Union.
It seems pretty clear to me that the Soviet engineers’ hard work on the Ryad hardware had ensured that copying IBM’s operating systems was much more straightforward to pull off, and saved a tremendous amount of time and effort that would otherwise have had to go into reverse engineering an IBM S/360 operating system with enough compatibility to run the billion plus dollars of S/360 applications that were available.
However I must emphasize that more was required than simply getting a tape with OS/360 on it to Moscow, inserting it into a Ryad system, and pressing the cyrillic equivalent of Ctrl-C and Ctrl-V. The Soviets most likely had to at least somewhat customize the IBM operating systems for each of the Ryad models, although obviously this was far easier than writing a compatible operating system from scratch. Clearly however the customization must have not been enormously extensive, or else so much of the English error codes and console commands etc would not have been carried over.
Full compatibility with IBM applications was still not perfected, with one of my sources stating that some modification of applications was not infrequently required, saying there would be hundreds of annoying but repairable problems. Other sources I found say that most applications ran without issue, so it's possible that S/360 applications generally ran fine, but required a decent amount of basic tweaking at a level that did not rise to a serious problem.
A declassified CIA briefing from the 1970s states that the East German made ES-1040 model did the best job of replicating the S/360 design and logic, and claimed that an unmodified IBM operating system was able to load and run on the ES-1040 without any issues, and had no issues running any IBM application that they tested.
I would say that in light of the state of Soviet hardware and technology in the 1970s, I wouldn’t be a bit surprised if they weren't able to exactly match the S/360 hardware specs, which would definitely require some tweaking or recompiling of applications in order to correct for differences in the bare metal.
Conclusions
The Ryad program was a hugely expensive undertaking whose final cost is probably not known. In 1967, the same year the program was announced, one Soviet source estimated that the program would cost 10 billion rubles, or about 13.8 billion dollars in 1967 American dollars. Contrast this with IBM’s expenditures of around 5 billion dollars to develop the entire S/360 line and it seems clear that Ryad from the get-go was expected to be hugely expensive. Some of this cost was spread around the other participating CEMA countries, but the bulk of it was of course born by the Soviet Union itself. Additionally, Ryad employed somewhere around 300,000 people in its development, while by comparison the entire US computer industry in 1972 only employed about 172,000.
However it should be noted that Soviet expenditures in any area are noticeably opaque, even now fifty years later, and a one-to-one correlation between dollars and rubles and the value gained by spending them is difficult to establish. The Ryad program may very well have cost quite a bit more than anticipated, which would honestly be the most likely outcome as large scale developments of new technology almost always are more expensive than predicted due to the large amount of unknowns.
Ultimately, for all its flaws and struggles, the Ryad was the first successful family of computers in the Soviet Union, and the first third generation system to enter significant serial production. It undoubtedly advanced the Soviet Union's computer capabilities significantly, and gave Soviet users the option to tap into the massive number of S/360 applications that had sprung up around IBM’s ecosystem.
As Ryad production ramped up throughout the 1970s, other competing or outdated systems began to wind their production down, with Ural and Minsk-32 machines ceasing production by 1975.
Production struggles continued to be endemic however. The Ninth Five Year Plan, which ran from 1971-1975, scaled back its projections for Ryad production to a mere 10-15 percent of the original target number, although CIA estimates of Ryad production do show a rapid growth in systems manufactured each year. However a lot of this increase in production was in lower end Ryad systems, not high end ones.
Ryad systems were not cheap, individual Ryad machines, especially in the early days, were actually far more expensive than other Soviet computers with comparable specs. The program had absorbed a massive amount of resources and talent, but I would say that the Ryad systems more or less met the requirements laid out for it in 1967. Soviet computer technology took a big step forward in both hardware and software, peripherals became far more widely available and of much better quality, and the increased amount of available Ryad systems helped build up Soviet computer science and create more engineers and programmers.
Software seems to have been the greatest beneficiary of the Ryad program, as one source says that by contrast with the Minsk-32, which was in serial production for 18 months before any real software was available for it, Ryad had a decent amount of programs available when it was still in prototype stage and well before it entered serial production. Although I do suspect that a lot of the software available was probably ported from the S/360, something the MINSK series could of course not make use of.
In spite of all the money and talent spent on it, the Ryad program did not manage to close the gap with American computer technology. What it primarily did was ensure that the Soviet Union didn’t lose a lot more ground, giving them the tools and technology needed to move into the third generation of computing instead of being stuck in the second. Soviet systems, even the best of the Ryad models that were in short supply, remained a number of years behind what was available in America. Additionally, while the Soviets had been working to bring the Ryad systems to production, the pace of technological achievement in the West had only speeded up.
As a CIA analysis from 1980, declassified in 2007, states in its abstract: while they were approximately seven years behind the West in the mid-1970s in general purpose computer products, they will be using general purpose computers in the late 1980s that are comparable to Western models of the early 1970s. This assessment of the lag in Soviet computer capabilities slightly varies from a different CIA assessment from 1977 that was declassified in 2001, which stated that the Soviet computer industry was a full ten years behind the West, but that may be just due to making a bit of a generalization versus a specific guess. Either way, both assessments were quite accurate as Soviet computing technology would remain around a decade behind the west in general and American in particular, all the way up to the collapse of the Soviet Union in 1991.
But none of this takes away from the fact that Ryad was a considerable achievement and a significant step forward for Soviet computing. The Ryad program cost billions of dollars, employed hundreds of thousands of people, who were all spread across a vast geographic area, and who all labored under the constraints of poor technology, poor communications, and part unavailability and shortages, as they struggled to adapt the S/360 architecture into a workable Soviet line of computers.
At the end of the day, the Ryad program heralded a new era in Soviet computing, and the best overall summation of its effect is probably this one, since 1968 a higher priority has been given to computer production and more effort has been put into coordination. A far smaller range of machines is produced in far greater numbers and there has been much greater emphasis on ensuring compatibility between different models and series. This effort has, at the least, prevented Soviet computer technology from slipping further behind American and created a more favorable outlook for the future.
Ryad was also the basis for a second generation family of systems, Ryad 2, that aimed to jump its IBM compatibility a whole generation, from the aging S/360 line to the newer S/370 line. But that's a story for another time.