The Politics of Networking a Nation

Eden Medina

In 1981, one year before his death, the Soviet cybernetician and computer pioneer Victor Glushkov published the book What Is the OGAS? OGAS was the Russian acronym for All-State Automated System for Gathering and Processing of Information for the Accounting, Planning, and Governance of the National Economy, USSR—a good illustration of the Soviet tendency toward lengthy names. The unlikely intended audience for the book: Soviet schoolchildren.

Since 1962 Glushkov had pushed in vain for the Soviet government to develop a national computer network capable of running its command economy. This network was supposed to improve Soviet economic management, encompass the entire Soviet Union by using existing and new telephone wires, and operate as a pyramidal hierarchy. Its base was to be 20,000 computer centers connected to 200 mid-level decision centers, then to a single central processing center in Moscow. These computers would use mathematical modeling to locate, rationalize, and improve inefficiencies. The system would also make all government documents electronic and permit economic workers to access them as needed.

But Glushkov’s 20 years of battling members of the Soviet government and military diminished the scope of his vision and the chances of its realization. As Benjamin Peters writes in his book How Not to Network a Nation: The Uneasy History of the Soviet Internet, Glushkov’s decision to publish a book about this computer network for a decidedly younger audience reflected “his sense of disappointment with his own generation.”

When a technology becomes part of daily life, its success conceals the alternative paths not taken. This is certainly true in the case of computer networks. We know that the Internet evolved from the US Department of Defense–funded network ARPANET, which was built to help scientists share resources and collaborate. Popular histories, science documentaries, and, most recently, a film by Werner Herzog have documented its rise as a global infrastructure.

Yet ideas about computer networking did not take root exclusively in the United States, nor did all networks resemble the ARPANET of the past or the Internet of today. When we look at networking projects that fell by the wayside—such as the OGAS—we learn not only about technological alternatives but also about the nations and institutions that created them. We also see how distinctive politics have shaped the form and function of technological systems, and not always in obvious ways.

<i>Victor Mikhailovich Glushkov</i>. Source: History of Computing in Ukraine,

Unlike the US foray into computer networking, the Soviet attempt never came to fruition. The historian of Soviet science Slava Gerovitch echoed Glushkov’s frustration more than two decades later when he described the Soviet networking experience as an “InterNyet.” However, the Soviets did pursue multiple efforts to build civilian computer networks, and these efforts reflected the context of their creation: they sought to rationalize the Soviet bureaucracy and were bound up with the larger socialist dream of using computers to manage the command economy.

As Peters shows, Cold War political goals such as improved economic management created the motivation for these systems, but political interests also thwarted their construction. For example, in 1959 the Soviet military scientist Anatoly Kitov proposed building a computer system that used existing military networks. However, his proposal criticized both military and civilian efficiency and thus led to his dismissal from the Soviet military.

In the case of the OGAS, competing interests within the Soviet bureaucracy also worked to counter attempts to rationalize economic management with computers. The system promised to improve the functioning of the state by eliminating corruption, identifying inefficiencies, increasing the amount of data collected and stored, and making information more readily available, all of which also served to implement new forms of state surveillance and control. At the same time, those who benefited from black markets, bribes, and bartering had little reason to support a technology aimed at their elimination. Nor did the different ministries, agencies, and offices in the Soviet state wish to adopt a plan that might limit their influence, reconfigure power relationships in ways that were disadvantageous, or cut their funding.

As a result, the government chipped away at Glushkov’s ambitious plan to create a national network of tens of thousands of computer centers across the Soviet nation and reduced it to the creation of hundreds of local computer centers during the 1960s and 1970s that were not even connected. Eventually, the Soviets abandoned the network project altogether. “If the Internet is not a thing but an agreement,” Peters writes, “perhaps the Soviet Internet is not a thing but a disagreement.” Peters argues that we have more to learn from these disagreements because of the competing interests they lay bare.

Competing interests within the Soviet bureaucracy worked to counter attempts to rationalize economic management with computers.   

Peters’s book performs a valuable service by making our understanding of computer networks more global and by deepening our understanding of how such networks emerged in the context of socialism. Nodes in comparative computer network history are few. In fact, I was asked to review Peters’s book because of my work on the history of a different socialist computing project, Cybersyn, which was built in Chile during the government of Salvador Allende. Like the Soviet OGAS, Project Cybersyn came into being as a means to manage the national economy and in response to its political context.

Like Peters, I find value in comparing the two networks as a way to unpack the observation that politics shape the design of technological systems while also illuminating the differences in attempts to engineer socialism during the Cold War. Socialist governments saw computer networks as a way to address a problem that was anathema in the US context: how to manage a command economy. Yet different approaches to socialism resulted in different network configurations. As Peters shows, Soviet networks were hierarchical. Glushkov’s plan for the OGAS consisted of thousands of data centers and collected extensive information about the economy and its workers.

In contrast, the design of Chile’s network reflected the country’s democratic road to socialist change. Allende gave priority to expanding the nationalized sector of the economy but remained committed to preserving democratic institutions and civil liberties. The network design illustrated this dual commitment by giving the state greater control of the economy while introducing mechanisms that preserved the autonomy of the factories within the state-run sector and increasing worker participation in decision-making. Instead of seeking to optimize performance, Cybersyn permitted a range of performance values. Factory managers could enjoy a certain amount of leeway in how they ran their factories so long as they satisfied production goals.

The Chilean example provides an additional point of comparison for understanding how political ideas and beliefs structure the design of computer networks. It also illustrates the centrality of the Cold War in the global history of computer networking. For example, Paul Baran at RAND envisioned building a communication system capable of surviving the effects of a nuclear war. He proposed routing message blocks of information across a distributed network, an insight that contributed to the packet-switching architecture of the ARPANET. As Peters shows, US innovations in computer networking pushed the Soviets to develop their own networking projects and ways to optimize the performance of a centralized socialist economy. Chile’s networking design reflected its decision to pursue a political third way between the superpowers and thus struck a balance between maintaining freedom and centralizing control.

Network types by Paul Baran, developer of the distributed communication network system, early 1960’s

Yet these comparative case studies also illustrate the imprecise relationship of ideology to the creation of technological systems, as well as their limited ability to uphold a set of principles. As Peters notes, the US government used state funding to heavily subsidize the creation of the ARPANET so that its scientists could collaborate more easily. In contrast, the Soviets developed their network in an environment that did not regulate conflicting interests and brought them into competition with one another. This leads Peters to the surprising conclusion that, in the process of network creation, “the capitalists behaved like socialists while the socialists behaved like capitalists.”

Network design and human behavior yield results that are equally complex and contradictory. The Soviet OGAS system was designed to eliminate corruption and increase the transparency of government activities by making it easier to access documents and data. However, as Peters shows, bargaining, favors, and the informal exercise of power were essential to making the rigid Soviet state function and ultimately worked against Glushkov’s ability to garner support for the OGAS project.

Similarly, the designers of the Chilean network envisioned Cybersyn as protecting factory autonomy and facilitating worker participation. The system would incorporate the knowledge of workers into the factory models, which would then be coded into Cybersyn’s software. This move would give the Chilean revolution a new kind of worker participation in factory management, but it also could have disempowered workers in the long term by making their knowledge part of a system designed to increase and maintain factory productivity.

The two cases further reveal the slippery relationship between system architecture and power. The Chilean and Soviet systems both aimed to decentralize decision-making. Cybersyn did so by protecting factory autonomy, improving information flows within the factory, and limiting the possibility for government functionaries to intervene on the shop floor. Yet the system channeled all the data it collected to one mainframe computer located in Santiago. Because the Chilean government could make only one computer available for the project, Cybersyn had a centralized design by necessity. The Cybersyn team therefore put in place a number of organizational checks to prevent the system from becoming abusive, but government officials could override them deliberately or by accident.

The Soviet decision to eschew the market in favor of the state cannot explain the inability of the Soviets to network their nation.

The Soviet OGAS, like Cybersyn, provided the government with a form of decentralized decision-making. As Peters writes, “From the outset, the OGAS Project sought to bring the economic bureaucracy online … allowing a decentralized remote access to all economic workers, and allowing decentralized access for controlling and optimizing the information in those documents.” This broad access, Peters argues, constitutes a response to criticism—including criticism from members of the Cybersyn team—that the Soviet system was overly centralized.

However, the Soviet case reveals that this framing of centralization as abusive and decentralization as preserving freedom is overly simplistic and even specious. We must consider the architecture of the OGAS within the context of its larger purpose: to collect and analyze data on sociopolitical and ideological processes as well as information about all aspects of the Soviet economy, including its workers and their activities. Part of the significance of the OGAS story, therefore, is that it shows the exertion of state power not only through top-down centralized control but also through distributed surveillance, the kind made possible by the broad accessibility of information on economic activity for those working in other areas of the state machine.

In fact, one of the key differences between the Soviet system and the Chilean system was not their degree of centralization or decentralization but rather the amount of data each system collected. According to one estimate, the Soviet system would have required monitoring 50 million variables. The Cybersyn system could handle only 10 to 12 variables per factory. Instead of collecting everything, Chile’s factory modelers tried to determine which indicators were the most important for the specific problem of economic management. Chile’s limited data collection reflected its technological capacity, which could not compare with that of the Soviet Union. But Chile’s inability to collect and aggregate millions of variables of production activity also resulted in design decisions that substantially limited the government’s ability to surveil the shop floor.

Comparative network histories can help us understand the relationship of network design to privacy and the public good in other ways. Peters’s conclusion is especially provocative here as it plays with the concepts of private and public to observe that the OGAS story is not one of private (market) versus public (state) interests. The Soviet decision to eschew the market in favor of the state cannot explain the inability of the Soviets to network their nation.

<i>The first domestic Soviet semiconductor general-purpose control computer, Dnepr, designed under the leadership of Victor Glushkov and Boris Malinovsky, Kyiv.</i> (c. 1961). Source: History of Computing in Ukraine,

Peters instead places the blame on a different idea of the private—the competing personal interests that made the Soviet state function. He writes that while black markets, bribes, bartering, favors, and the promise of extending one’s influence and power all played an important role in the machinations of the Soviet bureaucracy and the operation of the command economy, they were not factors that the OGAS project even thought to capture in its remaking of Soviet economic management. In fact, the rationalization and optimization the system sought to implement would, by definition, eliminate the corruption that made the Soviet economy function smoothly.

Taking the argument a step further, Peters contends that we should redraw the lines that define public and private more generally. Public interests would then refer to those of the broader community (the nature of which Peters does not specify), and private interests would refer to those of self-interested actors and institutions, including state agencies, corporations, and the military.

Thus framed, the OGAS story of party leaders and economic bureaucrats who put their interests above the health of the state and well-being of the economy looks a lot like the networked society of today. Peters points out that Google, Facebook, and the National Security Agency all have remade different parts of our network infrastructure in ways that further their private interests and without the regulation that would protect public interests. Surveillance, Peters writes, “is the massification of private attention and the antithesis of public attention.” Private self-interest flourishes in the absence of regulations that protect the public interest. Peters thus connects a different idea of the private—one that draws from Hannah Arendt’s articulation of oikos—to contemporary debates about privacy.

Peters does not suggest how to resolve this dilemma, however, other than by appealing to our political will to protect the public interest. Nor does he explain how we arrive at the public interest without traversing the negotiated space of private interests and the power and influence they wield.

Examining the shortcomings of other networked societies helps us see in sharper relief the shortcomings of our own networked world.

Nonetheless, the comparison of the OGAS to the NSA, Google, and Facebook should give all of us pause. If the differences between the OGAS and Cybersyn illustrate the murky line between centralized and decentralized architectures (and how both could result in abusive forms of surveillance), placing the OGAS in comparison with the Internet suggests that the distributed network configuration of today’s Internet does not automatically increase information freedom or flatten the ways that governments and businesses exert power and influence. Indeed, as recent controversies surrounding NSA surveillance capabilities or Google’s market dominance have shown, certain actors have been highly successful in making themselves a required point of passage for the flow of information and in harnessing this information for their own benefit. Networks in practice do not fall neatly into categories of distributed, centralized, or decentralized, nor is there a necessary correlation between liberty and distributed or decentralized network architectures. Computer history therefore reveals the messiness of these classifications and the difficulties that arise when they are invoked as simplistic explanations for the exercise of power and the preservation of freedom.

However, if Soviet computer networking provides an example of how our present-day networks might tend toward dystopia, perhaps the Chilean story provides a dose of optimism as well as a useful thought experiment. If the OGAS represents the possibility of a future of unlimited data collection driven by private interests, perhaps Cybersyn offers an illustration of how limiting data collection can prevent private interests from becoming too all-knowing and powerful. In 1970s Chile, technological limitations made it impossible to collect massive amounts of data on factory activities and forced discussions in the early phases of design about which data would be the most useful for achieving the goal at hand. This shows that it is possible to see discriminating data collection as a design feature, not a bug, and that this kind of thinking can result in highly innovative organizational and technical engineering. It also shows that extensive data collection is a design decision.

This is not to say that the Chilean system is beyond criticism or that it is immune to the same critique leveled at the OGAS. Neither the OGAS nor Cybersyn came to fruition or even came close to being implemented in the way their creators imagined. Despite their different data practices, both systems have been criticized as overly centralized and potentially abusive. Both were a product of their Cold War contexts and reflected the specific geopolitics of their day. Both took as given that more data would reveal truths about the world and necessarily lead to more informed decision-making, reduced bureaucratic inefficiency, and improved governance. And neither system was connected to the political realities that their governments faced. But these are also valuable historical lessons.

In this sense, both Cybersyn and the OGAS could be viewed as examples of how not to network a nation, especially in the 21st century. Yet examining the shortcomings of these other networked societies helps us see in sharper relief the shortcomings of our own networked world. Moreover, examining how these systems addressed the technological and political challenges of their times reveals the possibilities that exist in the paths not taken.