The Soviet internet

There is a new and intriguing book out by Benjamin Peters called How Not to Network a Nation: The Uneasy History of the Soviet Internet, which outlines exactly what it claims to.  Here is one introductory excerpt:

In late September 1970, a year after the ARPANET went online, the Soviet cyberneticist Viktor Glushkov boarded a train from Kiev to Moscow to attend what proved to be a fateful meeting for the future of what we might call the Soviet Internet.  On the windy morning of October 1, 1970, he met with members of the Politburo, the governing body of the Soviet state, around the long rectangular table on a red carpet in Stalin’s former office in the Kremlin.  The Politburo convened that day to hear Glushkov’s proposal and decide whether to build a massive nationwide computer network for citizen use — or what Glushkov called the All-State Automated System (OGAS, obshche-gosudarstvennyi avtomatizirovannaya system), the most ambitious computer network of its kind in the world at the time.  OGAS was to connect tens of thousands of computer centers and to manage and optimize in real time the communications between hundreds of thousands of workers, factory managers, and regional and national administrators.  The purpose of the OGAS Project was simple to state and grandiose to imagine: Glushkov sought to network and automatically manage the nation’s struggling command economy.

They failed!  The author blames this not on backward technology, but rather “entrenched bureaucratic corruption and conflicts of interest at the heart of the system…”

Anyone interested in the history of the internet, comparative systems, or the history of the Soviet Union should read this book.


I sometimes wonder what the relationship is between the abortive Soviet failures in network technology and the impressive success of the modern Israeli capitalist economy in network technology. It would be nice if some smart young fellows whose good ideas had been beaten down by Soviet stupidity in the 1970s wound up getting rich in Israel in the 2000s.

Well Sergey Brin's Father and Grandfather were Soviet mathematicians and computer scientists.

The point about Israel is that they wanted to be the Soviet Union. As long as the Labour Party was in power they tried to force Israelis into a Soviet-style centrally planned economy with a focus on heavy industry. As absurd as an Israeli steel industry is.

By the 1970s this was absurd and both parties essentially gave up. They let Jews be Jews rather than erzatz-Soviets (who probably really wanted to be Germans). Deregulation followed and Israel has done well. We are not allowed to talk about national character but Israel did not take well to the heavy industry Northern European model. They have taken well to the high tech sector.

There must be a lot of interesting stories about all the talent that was cooped up in the Soviet Union.

One story I read was about the Soviet mathematician, Petr Ufimtsev, who extended Maxwell's equations, which Lockheed used for the Stealth Fighter and Northrop Grumman for the Stealth Bomber. He eventually got a job teaching at UCLA in 1990, where he finally found out that his equations had spawned an industry in Southern California.

About a decade ago when my son was in junior high, I went to pick him up at his classmate's house, a blond kid who was always taking Ukrainian folk dancing classes. I had vaguely heard that his dad was an aerospace consultant. It turned out the family owned an entire canyon in the Hollywood Hills right above Toluca Lake. It had to be close to the best piece of land in the San Fernando Valley, a north-facing canyon full of oak trees only about a quarter of a mile above Ventura Blvd. Whatever know-how dad had walked out of the Soviet Union with, he must have made sure not to make Ufimtsev's mistake and not be in position to cash in on it.

"About a decade ago when my son was in junior high...

Steve, you once wrote that if you got to pick a career again you'd be a law school professor. I'm just curious, did you happen to nudge Sailer Jr., in the legal education direction?

How do you think the Soviets had all that smart talent to begin with? Native Russian Jews and the Germans that were captured during the war.

The Soviets had lots of talent due to the...size of the population. They were the largest European country in population terms. Lots of Jews, and ethnic Germans (which made up the majority of their scientists). So just the numbers game would imply that they would have a lot more scientific talent, than say, Bulgaria.

The influence of Ufimtsev work on stealth is greatly exaggerated, mainly due to internet rumors, and the secrecy around stealth technology itself.

Considering that both Lockheed and Boeing had figured out the properties of stealth at least a decade prior to his work:

Boeing's Quiet Bird from 1960

Lockheed's Oxcart from the 1960s:

Which later led to SR-71 which incorporated radar absorbing structures in its leading edges to scatter radar signals in a predictable fashion. Convair had Project Kingfish even earlier which also used the same ideas for radar signal deflection.

Ufimtsev work helped, undoubtedly, as a piece in the overall puzzle of how to design a low-observable aircraft. But it was just a piece of the puzzle.

Convair Project Kingfish from 1959:

Flat bottom, canted tails, hidden engine inlets, and those triangular shaped structures all along the leading edge of the plane are aimed at doing the same thing as the faceted design of the F-117 does: direct radar energy in a predictable fashion away from the receiver.

So the ideas of how to do it were there way before. Ufimtsev equations helped them perfect the shapes to specific frequency bands.

ISR: Interesting

The InterNyet


I didn't make it up. The author of the book mentions he borrowed it from a historian.

That's racist, Steve!

Love it. Slava Gerovitch gets all the credit. Also available: Nyetwork(ing), Nyetscape, etc.

The Soviets had 10s of thousands of computer centers, in 1970? If by computer they meant mechanical computers and water-powered machines, maybe.

The first Soviet computer was created in 1950. At the time, it was "the first universally programmable electronic computer in continental Europe."

"The first Soviet computer was created in 1950."

Still, tens of thousands of computer centers by 1970 does seem like quite a stretch.

After MESM in 1950, a substantial number of USSR's bigger computers were actually clones of Western machines, from IBM 360 clones to PDP and VAX clones all the way down to Sinclair clones.

The software was often ripped off as well. I remember my employer, a software company, getting support calls from the USSR in the '80s at sites where they had stolen the software to run on a pilfered hardware design and were brazen or clueless enough to call for free support on the stolen software.

No such concept as "stolen" in the USSR, comrade. Given that there was no such concept as "property" either.

Great leads. The author of the book here. Just to clarify: I do not mean to claim that there actually *were* tens of thousands of computer centers in 1970; only that building and connecting them was a main feature of the OGAS proposal. As it happens, Glushkov's ambitious network vision builds in part on the momentum of Sergei Lebedev's lab that built the MESM. It seems like most of the cloning begins in earnest after 1975.

Just so everyone knows, the five questions of journalism are who, what, wear, when and why. Some new school theorist are suggesting "why ask why"is the six question, but call me old fashion but i'm going to stick to the basics.

Certainly provides strong evidence for the bot theory.


Before Hillary, Libya was a tropical paradise. Now it's a desert wasteland!

Stein-Kodos 2016!

The internet as we know it today wouldn't be created in today's environment either.

Or if it was, it would be like Yik-Yak - banned on university campuses.

After all, I remember what the internet was like. Micro-aggressions was pretty much all it was good for before the browser came along and let people look at porn.

It is infrastructure, so it would be killed by a Republican committee chairman.

There's much more (inflation adjusted) federal R&D spending now than there was when the ARPANET was created -

That's a good graph, but it is research rather than infrastructure build-out.

What should we be building, and are we? Municipal WiFi?

The Soviets failed, but the French succeeded. Kind of. Few people even remember the Minitel system now, but back when most of the U.S. was unnetworked or at best on BBSs, I saw this remarkable little terminal in multiple ordinary french homes. My reaction was like Cowen's first ride on the DC metro. "It felt like the future!" But it failed fort many reasons including inability to break out much past its home market, and inability to keep up with the other major developments going on in the next decade or so. Maybe there could have been a Silicon Paree.

I remember Minitel rolling out nationwide in France in 1982.

It was kind of like Bloomberg Terminals, except they're still around, which I've never understood.

"Bloomberg Terminal" is now just a piece of software installed on a PC (they may support other operating systems, not sure) that comes with an optional special keyboard.

"The Soviets failed, but the French succeeded. Kind of. "

No, it didn't succeed. France closed it down in 2012. A government sponsored project limping along for 30 years and then being shut down is not a sign of success.

"mission accomplished"

Who would use the internet in 1970, and how would they use it? Bill Gates and Steve Jobs were 15 in 1970, the PC and Mac years from creation. Did the Soviet scientists in 1970 envision a PC in every home? Of course, what has made the internet a financial success is advertising, or revenues generated from advertising, advertising for products made the old-fashioned way, in the world of atoms to use Peter Thiel's metaphor. Today, the PC and Mac are giving way to the "smart" phone, which has the advantage of ubiquity, so (highly personalized) advertising can be transmitted to the consumer round the clock, not just when watching television or reading a magazine or newspaper, a much more effective method for communicating with the consumer. In their wildest dreams Soviet scientists could never have envisioned what the internet has become, a marketing device better suited for the "dumb" consumer than the "smart" scientist. The technology of Bill Gates and Steve Jobs may have made it possible, but it's marketing on the internet that has made the internet the generator of great wealth.

The Internet was a very good research tool that people later realized would work OK as a global communications system.

Well, according to Cowen's excerpt from the book, it was an "Internet" whose purpose was not to be a platform for consumers and businesses and meme-generation, but more as a control systems communication platform for the national economy. So really, more akin to the kind of communications network you'll find at the center of an industrial plant that connects all the hardware together with various operator stations and control rooms, scaled up to the national level.

In the west the telecommunication companies came up with X.25, which suited their needs until a public Internet rolled over the world.

Ah, interesting. Minitel used X.25

Precisely. Think an industrial control system for the national command economy. It's precisely that difference that makes the project so interesting for me, since it makes it possible to separate the global analysis of networks from the political, social, and economic values (democracy, consumers, memes, etc.) that generated the internet as we know it today.

A terminal in every home perhaps? When I was in graduate school not so long ago, you could still see wires sticking out of the walls in the department offices that people used to use to connect their terminals to the mainframe in the basement. The PC was a game-changer but it is easy to imagine in alternate path where the infrastructure for centralized, mainframe-based computing develops faster than the PC and people connect to the outside world using a dumb terminal connected to a community mainframe of some sort. Indeed, that seems to be what the French accomplished.

A Chromebook is a Minitel with better computation at both ends, faster communication. Different is scale rather than different in kind.

As described in the excerpt in Tyler's post, it sounds like what the Soviets envisioned was more or less an economy-wide ERP system. Real-time reporting of production, inventory, orders from one factory to another, etc. That does sound like something you'd logically try to develop if your mindset is running a centralized, planned economy.

The failure wouldn't have had anything to do with the socialist calculation problem, would it?

Rich, right on. (The author here.) The story *is* tied into the socialist calculation debates. Chapters 2 in particular takes up Soviet economics and the debates between orthodox, liberal, and cybernetic Soviet economists. Hayek lurks in the background of my discussion of their internal debates.

Reminds of Allende's ideas in Chile about the same time.....

Indeed, here Tabarrok's 2009 post on Allende's Cybersyn

And wikipedia's (note the Star Trek like control room!)

And here's Jacobin's wistful history, that sadly concludes technology alone will not create a better world.

I remember being in a museum looking at electromechanical pendulum clocks. The clocks were connected to a central clock (more precise) that sent time corrections to all other clocks over telegraph wires. This worked well for factories, banks but it became crucial to train operation. The message in this network was only "time", but the idea of a network that helps to coordinate human efforts has been floating around for more than a century.

Noah Smith made an opinion piece on this topic a couple days ago. Algorithms for optimizing communication and manage human other words: automated trading for everything.

Fascinating! Any more details about the electromechanical pendulum clock exhibit?

I am intrigued by how the high-frequency trade piece often leans on intensely mathematical visions of automated trading, as in this piece. I am also fascinated by the curious (but in no way precise) resonance between modern virtual banking and Glushkov's short-lived proposal in the early 1960s to replace hard currency with a system of electronic receipts (see chapter four and five for a bit more), or what we might see today as a predecessor to online banking. Glushkov foresaw the dissolution of hard currency as a fulfillment of Marxist prophecy, which is a fascinating contrast from bitcoin and post-fiat currency discussion today.

This excerpt has the Stalin Index of 1 (i.e. an irrelevant mention of Stalin in the very first paragraph.) This strongly suggests that the whole thing is most likely a piece of useless hackery. Then there is an amusing mistranslation of the system's name. "Obshche-gosudarstvennyi" actually means common or shared state (system) but here it's translated as "all-state" to make it sound 15% more sinister.

Incidentally, OGAS has a Wikipedia page (in Russian but not English.) It says that the development of the system had ultimately stopped not because of "entrenched bureaucratic corruption, etc." but due to the beginning of the market reforms.

Hi, the author of the book here. I appreciate the general skepticism level here. The popular commentary on the Soviet Union, economics, and technology certainly deserve that--and more.

The excerpted paragraph is the twelfth, not the very first, in the introduction. Chapters one and two go on at length to make clear what this book is about: it is not at all interested in Stalin. It *is* interested in understanding what comes after Stalin--and in particular the technocratically "neutral" discourse around economic control that tried to fill the political vacuum left by his passing. How to retain control without the controller, I think, is very much a live question today.

I too found "all-state" a strange translation at first, although, after months in archives and dozens of interviews in three languages with the original project promoters, I realized the best way to render the project in English is to use exactly the scientists' words in English, or namely what they called the "all-state" system. This is not sinister, although I cannot help but turn a bit pink at the thought! This is simply treating historical evidence fairly.

I think I won't touch the last proposition that we weigh a wikipedia comment over a scholarly book thesis.

Moreover, I am eager to leave the question of whether this book is in fact a "piece of useless hackery" to the evaluation of readers of the book, although I do wonder whether perhaps the initial comment doth protest too much?

Benjamin, thanks for the reply.

I didn't mean to imply that your book is hackery. It certainly looks worthwhile and I will probably take a look. I was referring to the way it was presented. Unfortunately, this kind of presentation is typical for materials about the old USSR or modern Russia, and I believe it prevents understanding.

This book looks great! Another book on a similar subject (cybernetics in socialist countries), also published by MIT Press, is "Cybernetic Revolutionaries: Technology and Politics in Allende's Chile" (2011), which seems like it would be an excellent companion to this book. It'd be quite interesting to compare and contrast the development and use of cybernetics technologies in these two different socialist countries.

Thanks, Arjun! Your radar is precise: Medina's work on the Chilean Cybersyn project--and Eden herself--has been formative in the making of this book. Add in the GDR, the Chinese, and many, many others, and there is lots of rich veins left for mining in the emerging field of comparative network studies, I think!

Francis Spufford's semi-fictional (and fascinating) book Red Plenty also covers some of this territory -- Glushkov is a major character.

Precisely, Fred!

If you have space for only one book on Soviet economics, read Spufford's Red Plenty: it's a delightful and elucidating miniature of the Soviet Union and its rolling heyday of economic promise. Not to be missed.

Not enough people note that the effective "creation" of the internet as we know it today is the convergence of the network infrastructure, with the software protocols for the Web, plus the commercial possibilities inherent in a world of growing computer use and widespread commercialization potential. A good network without the commercial public market for computers, software, profit seeking advertising, wide open international competition, and even foolhardy entrepreneurship in the 1980s and 1990s would not have provided anything close to what we think of as the Web today.

In Soviet Russia the internet searches you.

Terrific discussion! The author of the book and MR fan here. More replies in a moment, but first:

Some might enjoy an excerpt of the book in First Monday:
And here's a book review in Nature:

Enjoy! @bjpeters

Comments for this post are closed