Advent of Computing - Episode 9 - Networking for a Nuclear War, the Soviets

Episode Date: July 28, 2019

Often times people assume the US is the homeland of the internet. Funded by the US Department of Defence, the first attempts at a large-scale network were started during the height of the Cold War, an...d a large part of it's design was redundancy and robust-ness. Some of the researchers were quite frank about it's purpose: to create a network that could survive an upcoming nuclear war. This military-hardened infrastructure was known as ARPANET. But that's only part of the story, and the US wasn't the first to the party. The fact is, the internet was born during the Cold War. This was an era that saw huge advancements in science, both for better and for worse. The space race put humans on the moon, and the nuclear arms race put humans dangerously close to annihilation. So it should be no surprise that America's counterpart in this age, the Soviet Union, was working towards their own proto-internet.

Transcript
Discussion (0)
Starting point is 00:00:00 Why was the internet created? I think this is a surprisingly central question to understanding the 21st century. Was it done as some altruistic act for the betterment of mankind, or did it simply come into being for the profit of some long-lost corporation? In truth, I think its reason for existing is much more interesting and complex than you'd first think. Oftentimes, people assume that the U.S. is the homeland of the internet. Funded by the U.S. Department of Defense, the first attempts at a large-scale network were started during the height of the Cold War, and a large part of its design was redundancy and its ability to maintain robustness. Some of the researchers were quite frank about its purpose,
Starting point is 00:00:44 to create a network that could survive the oncoming nuclear war. The military hardened infrastructure would eventually be known as ARPANET. But that's only part of the story, and the US wasn't the first to the party. The fact is, the internet was born during the cold war. This was an era that saw huge advancements in science, both for better and for worse. The space race would put humans on the moon, and the nuclear arms race would put humans dangerously close to annihilating ourselves. So it should come as no surprise that America's counterpart during this conflict, the Soviet Union, was working towards their own proto-internet.
Starting point is 00:01:26 But if the Russians were first to the scene at creating an early internet, then why aren't they remembered as such? And why did the United States attempt with ARPANET turn into the internet we know today? Welcome back to Advent of Computing. I'm your host, Sean Hatz. This is episode 9, Networking for a Nuclear War, the Soviets. Now, I don't think that it's correct to point to any one event as the birth of the internet. Something that complicated doesn't have any one genesis point, especially in this case.
Starting point is 00:02:04 Something that complicated doesn't have any one genesis point, especially in this case. Oftentimes, folk will talk about ARPANET as the precursor to the internet, when in reality, a modern internet user wouldn't even recognize it. I think it's far more accurate to talk about early networking attempts like ARPANET as the birth of the technology that forms the infrastructure of the modern net. Now, to complicate matters even more, because that's something I like to do, there were wide-scale networks going back as far as the early 50s, but these were primarily one-off demonstrations or purely military. For the purposes of this series, I'm going to exclude military-only networks.
Starting point is 00:02:45 A big part of the modern internet is the idea of general use for data transfer, something that the earliest networks really lacked. Both the Soviets and Americans had functional missile defense networks established shortly after World War II ended. But, as I said before, they were purely military use, and they were very purpose-built. While these systems were impressive and warrant their own discussion, this episode isn't going to be the place for that. But sadly, even ruling out military networks won't clear the picture up completely. The Soviet Union, in particular, offers an interesting and rich history that runs parallel and at least partly isolated from American advancements in computing.
Starting point is 00:03:28 On top of that, a lot of Soviet history in general isn't well known outside of the former Eastern Bloc, simply due to the years of protection from the Iron Curtain. And it turns out that part of that obscure history includes attempts at a nationwide mixed-use civilian network nearly 10 years before any US-equivalent project would go public. However, by the time ARPANET did finally roll along with the full backing of the US government, any early Soviet attempts at widespread networking were totally eclipsed, and in the coming years ARPANET would evolve into the modern internet that we would now recognize. In this episode, we're going to examine the Soviet proto-internet projects,
Starting point is 00:04:12 ahead of a discussion of US's ARPANET next episode. By the time we're done, I want us to see how these attempts compare to one another, and where exactly the Soviet options fell short and the US options succeeded. And hopefully somewhere along the way, we'll start to see the beginnings of what would become the modern internet that we would now be familiar with. To start the story, we have to go a ways back into the depth of the Cold War, and we have to talk a little bit about economics. I'll try to keep it light though. So in the 1960s, communism around the world was facing a sort of crisis. A big part of keeping the Soviet Union stable was done via a planned economy. Now, this is a different economic model where the government
Starting point is 00:04:58 keeps tight control on production and distribution of goods and services. It's counter to the market economy that we have in the states, where forces of supply and demand drive everything. And surprisingly, this planned approach takes a lot of record-keeping and forecasting. As the 50s turned into the 60s, disaster started to appear on the horizon. Depleted from a world war and poor management decisions, the Soviet government was having issues keeping pace with the planned economy. To help manage planning, more and more bureaucrats at every level were joining the fold. But as with most issues,
Starting point is 00:05:37 just throwing more people power at the problem just makes things worse. It doesn't always fix the underlying issue. Luckily, innovation usually comes at the heels of disaster. And this was perfect timing for a new invention. Computers and automation in general had been looked at as a capitalistic encroachment during the Stalin era. But that attitude started softening in the early 50s. 1952 saw the first practical softening in the early 50s. 1952 saw the first practical computer built in the USSR, the vacuum tube powered BESM-1.
Starting point is 00:06:13 In the ensuing years, computing started to be adopted more and more, but mainly in the military. And by 1954, proposals to use computers to plan the Soviet economy started to be floated. Up until this point, the planned economy and the forecasts that were needed for it were all done by hand. The idea was to use computers to aid in processing economic models. But to do that, you have to have really powerful computers, you have to have a lot of information, and you have to really dial in your mathematical models. Luckily, computers are perfect for doing this, because as the country grew, it became impractical for more and more bureaucrats and number crunchers to do it by hand. Now, that's all well and good, but the economists and mathematicians that wanted to work on
Starting point is 00:07:01 planning and forecasting needed access to computers in the first place. In the late 50s, civilian economic planners started to get access to some calculating power as special built computer centers. But the construction of these centers took time, and the need for computing power would always outpace supply. Beyond that, creating new computer centers was expensive, not just in terms of the new computers, but the infrastructure needed to power, house, and control them. In 1959, Anatoly Kitov, while under the employ of the Soviet military, proposed an interesting solution. Kitov was what we would nowadays call a computer scientist, but in the 50s was known as a
Starting point is 00:07:45 cyberneticist. In fact, Kithov was one of the first cyberneticists in Russia and an early champion of widespread computer adoption in general. While working in a computer center under the Ministry of Defense, Kithov noticed how underutilized the massive military mainframes were. Military computer centers had large and powerful computers, and all the infrastructure needed to run them continuously. But despite the glut of power, these defense computers were only intensively used during the day, leaving them idle at night. If some of this existing computing power could be used, then a lot of money could be saved. power could be used, then a lot of money could be saved. Over a series of proposals, Kitthoff outlined a first of its kind, a mixed-use civilian and military network. Kitthoff's network,
Starting point is 00:08:33 called the Economic Automation Management System, or EASU, I know the abbreviations don't line up, that's due to the Russian to English translation mincing words. Networks in general weren't a totally new idea. The very military centers that Kitav planned to use were already part of a defense grid. However, there were some big differences in Kitav's proposal that make this an important step towards the modern internet. Firstly, it was planned to be mixed use. In other words, EASU would be shared between military and civilian users. This may seem like a minor difference, but this means that the new network would have to handle general purpose data transfer,
Starting point is 00:09:27 something that hadn't been done before at a large scale. The other facet of EASU was, well, automation. It's there in the name. Local terminals and computers and factories and other centers of industry would feed data into larger military computer centers. The central node would then handle all the number crunching and decision making, using economic models to create and update economic plans. It would essentially free up a lot of the bureaucrats that were currently manually working on new economic plans. The cost saving here is pretty clear. Using and expanding existing infrastructure already saves a good deal of money. Even better, the automation of EASU would eliminate a lot of jobs across the board in industry, government, and military. But therein lies the problem with Kithov's network. The board that was vetting Kithov's proposal was military,
Starting point is 00:10:15 and a lot of the people on the board would be made redundant if EASU were to be rolled out. Even more, the military wasn't very keen on sharing its infrastructure with civilians. They saw it as inefficient and, even worse, a possible security risk. EASU would never see the light of day. The review board was so against the plan that Kitav was kicked out of the military and the Communist Party. His final proposal was even burned in effigy. This effectively killed any chance of EASU or Kitov networking the nation. But this wouldn't be the last attempt. In the coming years, even more ambitious projects would come out of the woodwork. I want to say one more thing about EASU before moving on. It seems like, without actually using the words, Kitov's describing a
Starting point is 00:11:06 timesharing system. These were early solutions for sharing large and expensive mainframes between multiple users and applications. If that's the case, then EASU would have been on the very forefront of this technology. I think it's amazing as a case of parallel evolution, since the first American timesharing project, MIT's CTSS, was started in the same year as the EASU proposal. Just three years later, a new enterprising cyberneticist would try to network the Soviet Union once again, and this time, the plan would come purely from the civilian sector. And this time, the plan would come purely from the civilian sector. In 1962, Alexander Kerkevich penned a proposal for a new network called the Unified All-Government System for the Transmission of Information, or ESS. Once again, the abbreviation is off from the translation. While a civilian, Kerkevich was no stranger to the latest developments in the USSR. When he wrote the proposal, he was serving as the deputy chair
Starting point is 00:12:12 on the Council on Cybernetics, established in 59. The council was basically an offshoot of the Soviet Academy of Sciences that was working to further computing within the nation. While EASU was all about economic reforms via computing, ESS was to be much more far-reaching. The plan was to use existing infrastructure such as telephone and telegraph lines to create a unified, general-purpose civilian network. So far, that's sounding relatively modern. Where it diverges, though, is in the actual structure of the network. One of the key elements of modern internet design is that it's structured as a distributed network. That basically just means that there's no one computer that's a quote-unquote center of the internet. Instead, it's managed by a lot of different smaller
Starting point is 00:13:05 computers that work in tandem, and they supply redundancy in case part of the network goes down. ESS, however, was a totally different beast. It was a roughly decentralized network. If the mantra of a distributed network is that no node or computer is special, then you could sum up a decentralized network as some nodes are special. In the case of ESS, the control center of the network was in Moscow. That's where all traffic would have to be routed through and where the bulk of the computing power would be housed. Smaller computers spread around the Soviet Union would be used to connect up terminals and other access points. In this way, ESS as a network functions sort of like a set of tributaries that are feeding
Starting point is 00:13:55 into an increasingly larger system until you hit Moscow. But despite the technical differences between ESS and the current internet, there are a shocking number of modern ideas folded into the earlier system. EASU had been primarily a means to aid in economic planning, but ESS was a lot more ambitious. Kerkovic proposed that all existing communications, be that phone calls, telegraph, radio, or TV broadcasts, be digitized and sent over the network. This is a hugely forward-thinking idea that wouldn't be totally implemented until relatively recently. On the practical side, digitizing all data transfer would free up more existing infrastructure for the network. That means that ESS would be a truly general purpose network, it could send any type of data over it. And beyond that, ESS was planned to work
Starting point is 00:14:52 for large-scale operations of state function. Since all the data in the USSR would already be on the network, Karikovic reasoned that it would be easy to use it to automate decision making for the state. Just like with the earlier EASU, this would have reduced overhead by replacing government employees and streamlining the economy and the government at large. But there was a darker side. Since you have to remember, we're still in the Cold War era. The US and USSR both had defense networks at this point, so naturally the arms race would have to bleed into many new developments, and ESS was no exception. It was planned to have priority levels for data that was transferred on the network,
Starting point is 00:15:38 meaning that a higher priority message could be jammed through over low priority traffic. message could be jammed through over low-priority traffic. The use of this was very explicit, to send nuclear launch codes for the event of nuclear war with the United States. Ultimately, the proposal for ESS was well-received. Our man Alexander was appointed as the head of a new institute under the Academy of Sciences just to oversee the further development and creation of his nationwide network. It looked like things were finally on track, but fate would step in before any progress towards the future could be made. In 1965, Kerikovic died, and it appeared that the drive behind the ESS project went with him. Kerikovic had offered a shockingly insightful glimpse of the future while showing that the Soviet government was open to the idea of a widespread network.
Starting point is 00:16:31 Part of the government's acceptance of ESS came from it being proposed as separate to military infrastructure. The pattern of isolated systems is starting to develop here already, but we'll see that repeat itself later on. The next network that I want to talk about was perhaps the most fully realized, or at least the longest lived. Starting at near the same time as ESS was OGAS, the All State Automated System, and it was a mix of old and new ideas. The chief architect on the project was Viktor Glushkov, another early pioneer of Soviet cybernetics. Making a surprise reappearance was our man Anatoly Kitov. He was involved with OGOS in an advisory role.
Starting point is 00:17:20 So, what was OGOS, and how did it differ from earlier attempts to network the USSR? Well, as the name suggests, this is another system designed primarily for economic planning. The primary goal of the network was to automate the planned economy of the entire Soviet Union, like Kidov's earlier EASU. The vision was that via automation, a perfectly efficient economy could be created. And if you could create a perfect economy and automate it, then it would be a small step to creating a perfect government and then a perfect country. As I said earlier, the path to creating a better economic plan was economic modeling, something that had been attempted and was used on small
Starting point is 00:18:06 scales already by 62, but to properly model the entire Soviet Union, you end up needing information on everything that goes on within the Soviet Union. This is where networking comes into play. Lushkov proposed to network the entire country to collect and transfer data, just as others did before him. But there were some important differences in the details. First off, OGOS would be completely civilian, totally disconnected from any military infrastructure, similar to ESS. The idea of mixed civilian military networks was dead at this point. It was also planned to use existing telephone lines, instead of the take everything that exists and use it for networking approach of ESS.
Starting point is 00:18:54 The architecture of the network was also unique. While not fully distributed, OGOS was closer to decentralized, at least more decentralized than earlier attempts. This is definitely a step in the right direction. Blushkov designed a three-tier layout, with the top tier being a large computing center in Moscow. The second tier would be around 200 regional distribution centers dispersed around the country at key strategic or infrastructure points. the country at key strategic or infrastructure points. The bottom of the pyramid would be made up of 200,000 or more remote access points. These would be terminals or smaller home computers or computers in offices. In this configuration, any two computers could communicate with one another
Starting point is 00:19:40 without necessarily having to go through Moscow. And instead of having all data stored in one central data bank, information was to be spread across the entire network. Anyone with access to the network could access anything connected to the network. This may seem simple today since we just have to log into the internet and we can connect to anywhere in the world, but in the 60s this wasn't really a given for networking. The overall design was well-suited for large-scale data collection and storage, since you can scale it up by adding more computers to the network. But also, it had a few key implications.
Starting point is 00:20:18 Firstly, the system could still run if Moscow was cut off. The obvious application was, well, to keep the network hardened in case of an outside attack. The distribution of resources also meant that the seat of government wouldn't have to be in any one place. By digitizing state documents and communication, it would be simple to run meetings and coordinate the Soviet government from anywhere with a network connection. The other huge part of the conception of OGOS was local planning. By having regional computing centers with access to data from the entire network, OGOS would enable coordinating planning and managing local governments with the entire country at large. Part of this was creating large digital dossiers
Starting point is 00:21:06 of people of importance, both inside and outside the Soviet Union. This would end up being a sticking point for a lot of critics of the network. The combination of large-scale data collection and compiling personal dossiers was a scary step towards technologically-backed surveillance, and it's something that today we see all around us. Just really quick though, I want to touch on one other scarily modern proposal from OGOS. That's paperless currency. One of Glushkov's plans was to use OGOS to facilitate digital transactions, totally removing physical currency from the country.
Starting point is 00:21:46 This would be accomplished by a large-scale chain of transactional data. In other words, just keeping track of who spent what where and how money would change hands from transaction to transaction. Some of the details here are different, and there isn't really an implementation that we can point to, but the sound's an awful lot like Bitcoin or other blockchain-based digital currencies. So, back to the main point at hand. Overall, OGOS is another super forward-thinking proposal. It takes a lot of the good and the bad of previous Soviet attempts and it adds in some new features to create a network that's actually
Starting point is 00:22:25 pretty close in description to the modern internet. So how was the master plan received? Plushkov submitted his designs for OGOS in late 62. A few months later, in 63, the Soviet Politburo signed off on the plan. Over the next seven or so years, OGOS would run as a state-sponsored project, both politically and financially backed. Part of the reason this project was kept alive while early networks failed had to do with Glushkov's willingness to play politics. He was very much willing to change his plans to match with what the Soviet government and his critics wanted. to change his plans to match with what the Soviet government and his critics wanted. Sadly, that would also lead to part of the undoing of his project.
Starting point is 00:23:15 As OGOS started, it became apparent that its costs would be monumental. The network in total was projected to take somewhere in the neighborhood of 300,000 technicians working for 30 years before it could take its full form. And the cost? At least 20 billion rubles, or roughly 170 billion US dollars today. There was the expectation that the network would save the economy of the Soviet Union, but with this amount of resources having to be allocated to it for this amount of time, questions started to be raised. Unsurprisingly, as the OGOS project would roll along through the years, the estimated expenditures would start to be tweaked. Part of this was due to Glushkov's
Starting point is 00:23:58 willingness to change the designs to make the system more palatable to the administration. Ideas like paperless money were scrapped pretty early on, since that turned out to be too great of a change. Even while the network was being planned and details were being ironed out, the overall scale of the project would be radically cut back. A big factor in the final decline of OGOS was the factionalism inside the Soviet Union's government. final decline of OGOS was the factionalism inside the Soviet Union's government. In the case of EASU,
Starting point is 00:24:33 the cracks became immediately clear. The military would refuse to share its resources, so the project was scrapped immediately. But it was a little different with OGOS. The government at large agreed that a network was needed, and that automated coordination would be needed. A networked nation would be the best way to reform the economy into something more workable. However, to fully commit, each part of the government, each separate ministry, would need to give up some amount of control, and not everyone was as willing to do that as Glushkov hoped. By 1970, the final nail was driven in the coffin, and OGAS's funding was terminated.
Starting point is 00:25:10 In its place, a small patchwork of networks would start to form. Each separate ministry of the government would end up creating its own smaller network that operated just in their area of influence. Attempts to rekindle OGOS and any nationwide network project would continue to come up sporadically, but the fire had already gone out. In a matter of years, it simply became too expensive to try to replace all the smaller networks with a single unified network. And each separate internet was already too different to bridge into some nationwide computing resource.
Starting point is 00:25:47 So by the end of the 70s, it was really a foregone conclusion that the Soviet Union wouldn't be the heartland of the internet today. That wraps up the story for today. Ultimately, all attempts at a unified Soviet-wide network would fail. To me, at least, it seems that the culture in the Soviet Union's government was partly to blame. The country had an adversarial view of computers and automation during the Stalin years. This did loosen up as Khrushchev came to power in 1953, but the USSR was still a late player in the field of computing in general. Adding to this, proposals like OGOS and ESS were wide-sweeping reforms that would change
Starting point is 00:26:35 almost all aspects of Soviet life and restructure the government if they came into place. Ultimately, who knows if any of these projects would have been able to succeed even if they had been totally backed and had cooperation of the nation. In the next episode, we'll be looking at America's own early internet project, ARPANET. But there's definitely a lot more to be said about the Soviet proto-internet. If you want to read more on the topic, I'd highly suggest finding a copy of How Not to Network a Nation by Benjamin Peters. It covers a lot more context than I could get into here, and even better than that, it has citations, so you can dig deeper if you want to do that. Thanks for listening to Advent of Computing.
Starting point is 00:27:19 If you like the show, then take a minute to share it with your friends. We're growing a lot just via word of mouth. You can rate and review on Apple Podcasts. If you have any comments or suggestions for a future show, go ahead and shoot me a tweet. I'm at Advent of Comp on Twitter. And as always, have a great rest of your day.

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.