'darpa' Episodes

How Not To Network A Nation: The Russian Internet That Wasn't

     11/2/2020

I just finished reading a book by Ben Peters called How Not To Network A Nation: The Uneasy History of the Soviet Internet. The book is an amazing deep dive into the Soviet attempts to build a national information network primarily in the 60s. The book covers a lot of ground and has a lot of characters, although the most recurring is Viktor Glushkov, and if the protagonist isn’t the Russian scientific establishment, perhaps it is Viktor Glushkov. And if there’s a primary theme, it’s looking at why the Soviets were unable to build a data network that covered the Soviet Union, allowing the country to leverage computing at a micro and a macro scale 

The final chapter of the book is one of the best summaries and most insightful I’ve ever read on the history of computers. While he doesn’t directly connect the command and control heterarchy of the former Soviet Union to how many modern companies are run, he does identify a number of ways that the Russian scientists were almost more democratic, or at least in their zeal for a technocratic economy, than the US Military-Industrial-University complex of the 60s.  

The Sources and Bibliography is simply amazing. I wish I had time to read and listen and digest all of the information that went into the making if this amazing book. And the way he cites notes that build to conclusions. Just wow.

In a previous episode, we covered the memo, “Memorandum for Members and Affiliates of the Intergalactic Computer Network” - sent by JCR Licklider in 1963. This was where the US Advanced Research Projects Agency instigated a nationwide network for research. That network, called ARPAnet, would go online in 1969, and the findings would evolve and change hands when privatized into what we now call the Internet. We also covered the emergence of Cybernetics, which Norbert Wiener defined in 1948 as a the systems-based science of communication and automatic control systems - and we covered the other individuals influential in its development. 

It’s easy to draw a straight line between that line of thinking and the evolution that led to the ARPAnet. In his book, Peters shows how Glushkov uncovered cybernetics and came to the same conclusion that Licklider had, that the USSR needed a network that would link the nation. He was a communist and so the network would help automate the command economy of the growing Russian empire, an empire that would need more people managing it than there were people in Russia, if the bureaucracy continued to grow at a pace that was required to do the manual computing to get resources to factories and good to people. He had this epiphany after reading Wiener’s book on cybernetics - which had been hidden away from the Russian people as American propaganda. 

Glushkov’s contemporary, Anatoly Kitov had come to the same realization back in 1959. By 1958 the US had developed the Semi-Automatic Ground Environment, or SAGE. The last of that equipment went offline in 1984. The environment was a system of networked radar equipment that could be used as eyes in the sky to detect a Soviet attack. It was crazy to think about that a few years ago, but think today about a radar system capable of detecting influence in elections and maybe notsomuch any more. SAGE linked computers built by IBM. 

The Russians saw defense as cost prohibitive. Yet at Stalin’s orders they began to develop a network of radar sites in a network of sorts around Moscow in the early 50s, extending to Leningrad. They developed the BESM-1 mainframe in 1952 to 1953 and while Stalin was against computing and western cybernetic doctrine outside of the military, as in America, they were certainly linking sites to launch missiles. Lev Korolyov worked on BESM and then led the team to build the ballistic missile defense system. 

So it should come as no surprise that after a few years Soviet scientists like Glushkov and Kitov would look to apply military computing know-how to fields like running the economics of the country. 

Kitov had seen technology patterns before they came. He studied nuclear physics before World War II, then rocketry after the war, and he then went to the Ministry of Defence at Bureau No 245 to study computing. This is where he came in contact with Wiener’s book on Cybernetics in 1951, which had been banned in Russia at the time. Kitov would work on ballistic missiles and his reputation in the computing field would grow over the years. Kitov would end up with hundreds of computing engineers under his leadership, rising to the rank of Colonel in the military. 

By 1954 Kitov was tasked with creating the first computing center for the Ministry of Defence. They would take on the computing tasks for the military. He would oversee the development of the M-100 computer and the transition into transistorized computers. By 1956 he would write a book called “Electronic Digital Computers” and over time, his views on computers grew to include solving problems that went far beyond science and the military. Running company

Kitov came up with the Economic Automated Management System in 1959. This was denied because the military didn’t want to share their technology. Khrushchev sent Brezhnev, who was running the space program and an expert in all things tech, to meet with Kitov. Kitov was suggesting they use this powerful network of computer centers to run the economy when the Soviets were at peace and the military when they were at war. 

Kitov would ultimately realize that the communist party did not want to automate the economy. But his “Red Book” project would ultimately fizzle into one of reporting rather than command and control over the years. 

The easy answer as to why would be that Stalin had considered computers the tool of imperialists and that feeling continued with some in the communist party. The issues are much deeper than that though and go to the heart of communism. You see, while we want to think that communism is about the good of all, it is irrational to think that people will act ways in their own self-interest. Microeconomics and macroeconomics. And automating command certainly seems to reduce the power of those in power who see that command taken over by a machine. And so Kitov was expelled from the communist party and could no longer hold a command. 

Glushkov then came along recommending the National Automated System for Computation and Information Processing, or OGAS for short, in 1962. He had worked on computers in Kyiv and then moved to become the Director of the Computer Center in Ukraine at the Academy of Science. Being even more bullish on the rise of computing, Glushkov went further even added an electronic payment system on top of controlling a centrally planned economy. Computers were on the rise in various computer centers and other locations and it just made sense to connect them. And they did at small scales. 

As was done at MIT, Glushkov built a walled garden of researchers in his own secluded nerd-heaven. He too made a grand proposal. He too saw the command economy of the USSR as one that could be automated with a computer, much as many companies around the world were employing ERP solutions in the coming decades. 

The Glushkov proposal continued all the way to the top. They were able to show substantial return on investment yet the proposal to build OGAS was ultimately shot down in 1970 after years of development. While the Soviets were attempting to react to the development of the ARPAnet, they couldn’t get past infighting. The finance minister opposed it and flatly refused. There were concerns about which ministry the system would belong to and basically political infighting much as I’ve seen at many of the top companies in the world (and increasingly in the US government). 

A major thesis of the book is that the Soviet entrepreneurs trying to build the network acted more like capitalists than communists and Americans building our early networks acted more like socialists than capitalists. This isn’t about individual financial gains though. Glushkov and Kitov in fact saw how computing could automate the economy to benefit everyone. But a point that Peters makes in the book is centered around informal financial networks. Peters points out that Blat, the informal trading of favors that we might call a black market or corruption, was common place. An example he uses in the book is that if a factory performs at 101% of expected production the manager can just slide under the radar. But if they perform at 120% then those gains will be expected permanently and if they ever dip below the expected productivity, they might meet a poor fate. Thus Blat provides a way to trade goods informally and keep the status quo. A computer doing daily reports would make this kind of flying under the radar of Gosplan, or the Soviet State Planning Committee difficult. Thus factory bosses would likely inaccurately enter information into computers and further the Tolchachs, or pushers, of Blat. 

A couple of points I’d love to add onto those Peters made, which wouldn’t be obvious without that amazing last paragraph in the book. The first is that I’ve never read Bush, Licklider, or any of the early pioneers claim computers should run a macroeconomy. The closest thing that could run a capitalist economy. And the New York Stock Exchange would begin the process of going digital in 1966 when the Dow was at 990. The Dow sat at about that same place until 1982. Can you imagine that these days? Things looked bad when it dropped to 18,500. And the The London Stock Exchange held out going digital until 1986 - just a few years after the dow finally moved over a thousand. Think about that as it hovers around $26,000 today. And look at the companies and imagine which could get by without computers running their company - much less which are computer companies. There are 2 to 6 billion trades a day. It would probably take more than the population of Russia just to push those numbers if it all weren’t digital. In fact now, there’s an app (or a lot of apps) for that. But the point is, going back to Bush’s Memex, computers were to aid in human decision making. In a world with an exploding amount of data about every domain, Bush had prophesied the Memex would help connect us to data and help us to do more. That underlying tenant infected everyone that read his article and is something I think of every time I evaluate an investment thesis based on automation. 

There’s another point I’d like to add to this most excellent book. Computers developed in the US were increasingly general purpose and democratized. This led to innovative new applications just popping up and changing the world, like spreadsheets and word processors. Innovators weren’t just taking a factory “online” to track the number of widgets sold and deploying ICBMs - they were foundations for building anything a young developer wanted to build. The uses in education with PLATO, in creativity with Sketchpad, in general purpose languages and operating systems, in early online communities with mail and bulletin boards, in the democratization of the computer itself with the rise of the pc and the rapid proliferation with the introduction of games, and then the democratization of raw information with the rise of gopher and the web and search engines. Miniaturized and in our pockets, those are the building blocks of modern society. And the word democratization to me means a lot.

But as Peters points out, sometimes the Capitalists act like Communists. Today we close down access to various parts of those devices by the developers in order to protect people. I guess the difference is now we can build our own but since so many of us do that at #dayjob we just want the phone to order us dinner. Such is life and OODA loops.

In retrospect, it’s easy to see how technological determinism would lead to global information networks. It’s easy to see electronic banking and commerce and that people would pay for goods in apps. As the Amazon stock soars over $3,000 and what Jack Ma has done with Alibaba and the empires built by the technopolies at Amazon, Apple, Microsoft, and dozens of others. In retrospect, it’s easy to see the productivity gains. But at the time, it was hard to see the forest through the trees. The infighting got in the way. The turf-building. The potential of a bullet in the head from your contemporaries when they get in power can do that I guess. 

And so the networks failed to be developed in the USSR and ARPAnet would be transferred to the National Science Foundation in 1985, and the other nets would grow until it was all privatized into the network we call the Internet today, around the same time the Soviet Union was dissolved. As we covered in the episode on the history of computing in Poland, empires simply grow beyond the communications mediums available at the time. By the fall of the Soviet Union, US organizations were networking in a build up from early adopters, who made great gains in productivity increases and signaled the chasm crossing that was the merging of the nets into the Internet. And people were using modems to connect to message boards and work with data remotely. Ironically, that merged Internet that China has splinterneted and that Russia seems poised to splinter further. But just as hiding Wiener’s cybernetics book from the Russian people slowed technological determinism in that country, cutting various parts of the Internet off in Russia will slow progress if it happens.

The Soviets did great work on macro and micro economic tracking and modeling under Glushkov and Kitov. Understanding what you have and how data and products flow is one key aspect of automation. And sometimes even more important in helping humans make better-informed decisions. Chile tried something similar in 1973 under Salvador Allende, but that system failed as well. 

And there’s a lot to digest in this story. But that word progress is important. Let’s say that Russian or Chinese crackers steal military-grade technology from US or European firms. Yes, they get the tech, but not the underlying principals that led to the development of that technology. Just as the US and partners don’t proliferate all of their ideas and ideals by restricting the proliferation of that technology in foreign markets. Phil Zimmerman opened floodgates when he printed the PGP source code to enable the export of military-grade encryption. The privacy gained in foreign theaters contributed to greater freedoms around the world. And crime. But crime will happen in an oppressive regime just as it will in one espousing freedom. 

So for you hackers tuning in - whether you’re building apps, hacking business, or reingineering for a better tomorrow: next time you’re sitting in a meeting and progress is being smothered at work or next time you see progress being suffocated by a government, remember that those who you think are trying to hold you back either don’t see what you see, are trying to protect their own power, or they might just be trying to keep progress from outpacing what their constituents are ready for. And maybe those are sometimes the same thing, just from a different perspective. Because go fast at all costs not only leaves people behind but sometimes doesn’t build a better mousetrap than what we have today. Or, go too fast and like Kitov you get stripped of your command. No matter how much of a genius you, or your contemporary Glushkov are. The YouTube video called “Internet of Colonel Kitov” has a great quote: “pioneers are recognized by the arrows sticking out of their backs.” But hey, at least history was on their side! 

Thank you for tuning in to the History of Computing Podcast. We are so, so, so lucky to have you. Have a great day and I hope you too are on the right side of history!


(OldComputerPods) ©Sean Haas, 2020