In the early sixties, a mathematician from Kyiv proposed linking the entire Soviet Union with a computing network — and spent the next twenty years watching the state he had designed it for do everything possible to ensure it never happened.
Before Viktor Glushkov formulated his project, another man had made the first attempt. In 1959, military engineer Anatoly Kitov sent the country’s leadership a proposal — to combine the computing power of military and civilian agencies into a single network for managing the economy. The idea was rational in an engineer’s way: why keep expensive machines separate if, at night, when no military calculations were being run, those same computers could process economic tasks? Kitov framed it as the “Red Book” project and addressed it directly to Khrushchev, bypassing the intermediate authorities. The military establishment saw the proposal as an encroachment on its resources. The initiative was rejected, Kitov was expelled from the Party and removed from his post. The lesson was clear: anyone proposing to share computing power risked being left without any of his own.
This episode set the pattern for all subsequent attempts. Any network, by definition, requires its nodes to exchange data, and in a system where information is the currency of power, data exchange was seen as a threat. Kitov was the first to step on that mine. Glushkov would approach the same minefield three years later, but from another flank and with a far more ambitious design.
In 1962, Academician Glushkov, by then already director of the Institute of Cybernetics in Kyiv, presented the OGAS project — the Nationwide Automated System for Accounting and Information Processing. The plan had three tiers: one central computing hub in Moscow, up to 200 regional centers across the country, and up to 20,000 local terminals at enterprises. Data on production, inventories, and deliveries was to flow from the bottom up, be processed, then return as adjusted plans — in real time.
Put crudely, Glushkov wanted to turn the Soviet economy into an organism with a functioning nervous system. Gosplan would receive not reports two months out of date, compiled with allowances for inflated figures, but a live stream of numbers: how much steel had been smelted today, how many railcars were sitting idle in a Novosibirsk sorting yard, how much cotton had failed to reach a textile mill in Ivanovo. The system was meant to support planning, not replace planners — but the mere fact that data would no longer be filtered through dozens of bureaucratic hands changed everything.
The scale of the infrastructure required was daunting. The country needed thousands of computers compatible with one another — yet Soviet computing technology at the time was produced in scattered series, often incompatible even within a single factory. It needed communications channels capable of transmitting data between regions — yet the telephone network in rural areas could barely handle voice calls. It needed programmers, engineers, operators — hundreds of thousands of people who still had to be trained. Glushkov understood all this and proposed a phased rollout, but even the first phase required a political decision that never came before 1970.
In 1970, Glushkov submitted an official funding request. The figures he named silenced even the sympathetic: 20 billion rubles and a staff of roughly 300,000 people. For comparison, the USSR’s entire annual defense budget during that period was estimated at comparable sums. Glushkov argued that the system would pay for itself within a matter of years — the losses from inefficient planning, by his calculations, exceeded the cost of OGAS many times over. But budget logic and political logic are different disciplines.
The request was rejected. Formally — because of the cost. In reality — because by then a coalition had formed against the project that no scientist, not even one with academician status, could have pushed through. The key figure in the resistance was Finance Minister Vasily Garbuzov. From a departmental standpoint, his position was understandable: OGAS required a unified data format and transparency in financial flows, and that threatened the Finance Ministry’s monopoly on economic statistics. When every factory can send its numbers directly to the center, the intermediary becomes unnecessary — and the intermediary understood that perfectly well.
One detail gets me: Glushkov was not some naïve dreamer locked away in a laboratory. He attended meetings, negotiated, tried to find compromises with ministries. He proposed modular implementation — start with a few sectors, show results, expand. But every compromise diluted the very idea of a unified network. A sectoral ministry would agree to automation — but only its own, only under its own control, only with its own standards. A network without a common protocol is a collection of islands linked by ferries that run on schedules convenient to each island separately.
Alongside OGAS, the Soviet Union really was creating automated management systems — ACS. By the mid-seventies, there were hundreds of them. Individual enterprises, sectors, and agencies acquired their own computing centers, wrote their own software, solved their own problems. A metallurgical combine calculated the workload of its rolling mills. An oil refinery optimized its use of raw materials. Railway workers tried to track their rolling stock.
The problem was that every ACS existed inside its own information bubble. Data from one ministry’s system could not be passed directly into another’s: different formats, different machines, different programming languages, different ideas of what should count as a unit of record in the first place. When Gossnab said “a ton of rolled steel,” it meant one thing; when that same rolled steel passed through the system of the Ministry of Ferrous Metallurgy — it meant something slightly different. This mismatch at the level of definitions made integration technically agonizing even where there was enough political will.
There were alternative projects with similar ambitions. SOFE — the system for the optimal functioning of the socialist economy — proposed mathematically calculating optimal prices and resource allocation. EGSVC — the Unified State Network of Computing Centers — was an attempt to at least standardize the technical base. Akademset linked research institutions. Each of these projects addressed its own fragment of the problem, but none aspired to the wholeness Glushkov had built into his design.
In recent years, the project has accumulated a mythology that makes its real significance harder to grasp. OGAS is often called the “Soviet internet,” and the phrase is catchy, but inaccurate to the point of distortion. The internet — ARPANET, and then TCP/IP — was built as a distributed packet network without a single center, resilient to damage, designed from the outset for horizontal interaction among equal nodes. OGAS was hierarchical: data rose from terminals to regional centers, from regional centers to the Moscow hub. It was a tree, not a web.
Besides, the internet ultimately became a medium for people to communicate — email, forums, web pages. OGAS was designed for the exchange of economic data between institutions. No personal messages, no free access to information for citizens. Confusing it with the internet is like calling Ford’s assembly line a prototype of the autobahn: both involve movement, but their purposes are fundamentally different.
Still, the comparison is not meaningless in one respect: both OGAS and ARPANET emerged in the same era, from the same realization — that computing machines connected by communications channels were capable of something qualitatively different from isolated computers. The difference was that the American project grew out of an academic environment where horizontal exchange was the norm, while the Soviet one grew from a planning hierarchy where any unplanned exchange looked suspicious.
It is customary to say that bureaucrats killed OGAS. That is true — but not the whole truth. Even if Garbuzov and his like-minded colleagues had magically disappeared, the project would have run into material constraints that no Council of Ministers decree could overcome.
By the early seventies, the Soviet electronics industry lagged a generation behind the West, and in some fields — two. Mass production of reliable, mutually compatible machines was never established at the scale OGAS required. The decision to copy the architecture of IBM System/360 — the ES EVM series, launched into production in 1971 — was meant to solve the compatibility problem, but created a new one: dependence on someone else’s standards and chronic lag in mastering each successive generation. Communications channels remained a bottleneck: the long-distance telephone network was overloaded, while dedicated data transmission lines existed only in fragments.
But the most acute shortage was immaterial. OGAS required enterprises to transmit accurate data upward. Yet the entire logic of Soviet economic life pushed in the opposite direction. An understated target is easier to fulfill. An inflated report earns a bonus. Hidden reserves are insurance against unpredictable directives. A factory director who honestly reported to the system that three thousand tons of unaccounted-for metal were sitting in his warehouse risked losing that metal — it would immediately be reallocated. A system built on transparency required trust — and trust between the levels of the Soviet managerial hierarchy was a resource scarcer than semiconductors.
I think that is the central paradox: OGAS was needed precisely because economic data was unreliable — but it could function only with reliable data.
After the funding rejection in 1970, Glushkov did not stop. He continued developing the theoretical foundation, worked on individual elements of the system, wrote books and articles, spoke at meetings. The Institute of Cybernetics in Kyiv remained one of the country’s strongest research centers. But OGAS as a unified project ceased to exist in the political sphere — it turned into a set of scattered initiatives, each solving a local problem and none claiming Union-wide scope.
The seventies and early eighties became a time of quiet fading for the big idea. Sectoral ACS systems continued to proliferate, but without a common standard and without connectivity between them. Every ministry built its own digital fence. Gosplan still worked with paper balance sheets that no longer reflected reality by the time they were approved. The economy grew more complex — the range of products expanded, logistics chains grew longer — while the instruments of management remained essentially the same as they had been in the fifties.
Viktor Glushkov died on January 30, 1982, at the age of fifty-eight. With his death, the project lost the only person who possessed, all at once, the scientific authority, political connections, and stubbornness needed to raise the issue at the state level again and again. After 1982, no one spoke seriously of OGAS as a coherent undertaking.
Could OGAS have saved the planned economy? The question sounds dramatic, but at its core it is wrong — because it assumes that the problem of the planned economy was computational. Glushkov, for all his genius, started from the premise that a plan could be made precise if enough data were collected and processed quickly enough. But the plan in the Soviet economy was not a mathematical problem — it was the result of bargaining. Ministries bargained for resources, enterprises for targets, regions for investment. A computer could speed up the arithmetic, but it could not eliminate the bargaining — because bargaining was not a bug in the system but its load-bearing structure.
At the same time, dismissing OGAS as a utopia would be dishonest. Glushkov saw a real problem: by the sixties, the Soviet economy was producing more information than it could digest. The number of product categories ran into the millions, intersectoral connections into the tens of millions. Calculating the balance of such a system by hand was physically impossible, and Gosplan managed only through crude aggregation, which inevitably produced shortages and gluts. A computing network really could have softened this problem — if the data entering it had been honest.
The story of OGAS is the story of a technical solution colliding with an institutional reality in which data transparency was not a benefit but a threat. The ministries were not irrational in their resistance — they were defending their authority, as any bureaucratic structure does in any country. The difference is that in a market economy, prices do some of the coordinating work for which Glushkov designed his network — and do it without anyone’s permission. In a planned economy, there was no substitute for the price signal, and OGAS might have become one — but for that, it needed to receive what no one could give it: truth from bottom to top.
The project was never deployed as a unified Union-wide system. Some of its elements dissolved into departmental ACS systems, while the idea of a coherent computing network for managing the economy outlived both its author and the country for which it was created — but only as a subject of historical research, not as an engineering plan.