Rockets, Engines, and IoT: Lessons Learned from the USSR Moon Landing Program

“…I believe that this nation should commit itself to achieving the goal, before this decade is out, of landing a man on the Moon and returning him safely to the Earth…” John F. Kennedy, Special Message to Congress on Urgent National Needs, May 25, 1961

After John F. Kennedy had made this statement, the race to land a man on the Moon has begun. The United States had lost the first two rounds of the Space Race (the first artificial satellite and the first man in space) and was determined to fight back. Their only rival at the time, the Soviet Union, accepted the challenge.

Hundreds of thousands of scientists and engineers on both sides of the Iron Curtain got a new, grand mission. Leaving aside the political battle between communism and the free-market economy, it was an endeavor that drafted the most prominent minds of aerospace and computer science at that time.

Even 50 years later, we still reap the fruits of that battle. The development of integrated circuits (the predecessors of computer processors) was stimulated by the requirements of spacecraft flight computers.

If you’re reading this on an electronic device, you’re benefiting from the technological legacy of the Moon race.

The US Moon landing program was entitled Apollo, and its history is well-described. The history of the Soviet part of the Moon race, however, is far less known, but no less dramatical. As the United States had won the race and landed a man on a Moon first, the Soviet Union decided to cancel its Moon landing program and insisted that the USSR should never attempt to land a man on Moon. Let’s dive into that story and find out how we can relate the problems of the rocket scientists from the 1960s to modern technological challenges.

The USSR part of the Moon Race

Clash of egos and the ultimate compromise

The cornerstone of the Soviet Moon program was the N-1 rocket. The rocket’s chief designer Sergiy Korolev, the genius behind the previous Soviet space victories, was facing the paramount of engineering problems.

The super-heavy rocket was initially designed for other purposes, the funds for the development program were limited, and the Soviet Union had entered the Moon race too late. On top of that, Korolev conflicted with the most prominent Soviet rocket engine constructor Valentin Glushko.

Officially, they disagreed on the types of engines to be used for the N-1 rocket. Korolev insisted on using the kerosene/LOX fuel while Glushko opposed that idea, insisting on the UDMH/N2O4 propellants. Unofficially, the nature of their conflict was personal and dated back to Korolev's imprisonment in the Gulag labor camp.  

When Glushko had finally refused to design engines for N-1, Korolev turned to another Soviet engine designer, Nikolay Kuznetsov. Kuznetsov had previously designed only aircraft engines and was a novice in rocket engines. So, they reached a compromise: to use a cluster of smaller, low-thrust closed-cycle kerosene/LOX engines instead of several very powerful engines

Kuznetsov’s design bureau was assigned the development of this new fuel-efficient engine. The final design of the first stage of the rocket included 30 (!) engines placed in two nested rings. In total, there were 42 engines in all three rocket stages. At the same time, the US Saturn V rocket that delivered a man to the Moon had a total of 11 engines.

Nobody had attempted to design a rocket design that complex before, and the engineering risks turned out to be too high for the Soviet Union.

Sacrifice an engine, save a rocket

The early versions of Kuznetsov’s engines weren’t reliable. They didn’t just fail: the engines tended to explode. An engine explosion during a flight would destroy the nearby engines and the entire rocket as a result. The designers couldn’t figure out why the engines would explode, but they could specify a set of parameters describing a well-functioning engine.

To prevent explosions, they offered to monitor certain parameters of all engines’ performance closely and shut down the one that looked suspicious. That way, the engines would never reach a dangerous point and avoid explosions.

To compensate the difference in thrust, the engine on the opposite side of the “ring” would also have to be shut down. The rocket had an excessive thrust, and it could continue its mission with two pairs of first-stage engines shut down in-flight.

The KORD system

Each of the 42 engines had its own monitoring equipment that consisted of:

• primary sensors
• an electronic block of amplifiers
• a communication line
• a system control block linked with the engine control automatics system

The system was designed with a reaction time of 4–5 milliseconds — a time short enough to prevent the engine from falling into the dangerous working mode. They named the system “KORD,” an acronym for “Control (of) Rocket Engines” in Russian.

Korolev dies

In 1966, Sergey Korolev died at the age of 59 during a minor surgery. The details of his death are still unclear. His health was undermined by a 6-year imprisonment, part of which he had spent in a gold mine in the Kolyma region, the far East of Siberia. While being interrogated and tortured, he got his jaw broken. It never healed properly. Rumor has it that his broken jaw could have caused the fatality. In the process of the surgery, Korolev started to bleed. Doctors couldn’t intubate him because his broken jaws just couldn’t open wide enough. His weak heart was also to blame.

The Moon program continued, but the N-1 rocket program was derailed badly, and its outcomes got even grimmer.

KORD goes mad

The first launch of the N-1 rocket on February 21, 1969, was a failure.

Sixty-eight seconds after the launch, the KORD system issued a command to shut down all engines of the first stage. The rocket crashed 52 km away from the launch pad. A subsequent investigation discovered that the KORD system was susceptible to interference from a power generator. Besides, the KORD’s sensors picked an electrical short circuit in the power lines and interpreted it as a problem with the turbopumps. Following its algorithm, KORD shut down all engines, causing the rocket to crash.

Could the rocket continue its flight if the KORD command hadn’t been issued? We’re only left to guess. The short circuit was caused by a fire in the tail section of the first stage, and nobody knows how damaging it could have been at the later stages of the flight. Officially, KORD was named the root cause of the first N-1 launch failure.

Ironically, the system designed to save a rocket was the cause of failure during the very first flight.

The industrial revolution begins

The ideas employed by the designers of the KORD system weren’t unique. The same line of reasoning was happening all over the world, with automation and computerization as the main ideas. On the other side of the Iron Curtain, the Third Industrial Revolution was about to begin.

ARPANET (the predecessor of the Internet) was launched in 1969. Intel started developing the 4004 microchip (the grandfather of the modern computer processors) the same year. Computers became in charge of all critical industrial processes, and digital systems replaced the analogous ones. The new “automation” approach was adopted by various industries: aeronautics, power production, oil and gas, mining. Several years later, factory automation began: the SCADA systems date back to early ‘70s.

In the western world, the technologies that have previously been available to the military and space programs only, have quickly become cheap and proliferated on other industries. The 1980s was the era of microcomputers, and since then, digitalization took the world by storm.

IoT lessons learned 50 years later

Korolev and his KORD system engineers were solving a problem that can be related to modern IoT projects. They had pieces of machinery they wanted to control, a set of sensors to provide measurements, a central control unit that could make decisions and a feedback loop to execute commands over the machinery.

The analogy, though, might be rather weak. With all due respect for rocket engineering pioneers, the modern IoT projects have an extra set of challenges.

• Security was not an issue. In 1969, nobody wanted to hack the rocket in-flight, steal data from it or disrupt its functions.
• No need to manage the lifecycle of rocket components. You didn’t have to update the firmware, ensure a smooth transition between versions, patch vulnerabilities. The intended lifetime of a rocket was measured in minutes, and there was no scenario for reusing its parts.
• Interoperability and integration with other systems weren’t required. There was no need to support various network protocols and industry standards.
• The competitors weren’t reverse-engineering their hardware and software the very minute they released it. Probably, because the creators were intentionally keeping it a secret.

(Not) reinventing the wheel

Re-using the working solutions and not reinventing the wheel is the foundation of good engineering. The engineering legacy and spin-off technologies of the US Apollo program were used in future space and military programs and eventually spread into the civil use.

Similarly, in every IoT project, there are common parts, that can be reused and should not be developed from scratch. Commercially available IoT “platforms” offer exactly that: reusable tools and technologies. Various vendors provide IoT platforms, including the old-school players of the traditional automation industry and the new players from cloud computing.

AWS IoT in an excellent example of a commercial IoT platform that can work in near real-time conditions. With sufficient effort, you can now reproduce solution similar to KORD with a fraction of the cost and time required.

In tests we trust

One of the reasons the launch of N-1 failed was the lack of testing racks. The engineers couldn’t do a full-scale testing cycle with all engines working in all modes.

Predicting how a system with multiple connected devices may perform under various loads is not a trivial task. Today, the developers of IoT systems face the challenge of ensuring scalability: the number of connected devices can be virtually unlimited. Simulating a large pool of connected devices is a good way to test back-end services and their performance.

We’re lucky to be using elastic cloud services that allow us to simulate a large number of connected devices without excessive costs. Check out this device simulator from AWS 

The fate of the Soviet Moon Program

Korolev died before the first launch of the N-1 rocket. Three years after his death, the US landed on Moon in 1969, meaning that the USSR has officially lost the Moon Race. The development of the N-1 rocket continued until 1974, with a total of 4 failed flights.

Ironically, it was Valentin Glushko, who canceled the N-1 program. Glushko was appointed the chief of a new design bureau that has been merged with Korolev’s one. About 150 remaining rocket engines manufactured for the N-1 rocket were preserved. After the fall of the Soviet Union, some of those engines were used in the United States for commercial launch services. The entire Soviet Moon program was kept secret until 1989.

As the United States had won the race and landed a man on a Moon first, the Soviet Union decided to cancel its Moon landing program and insisted that the USSR should never attempt to land a man on Moon.

Lessons applicable to IoT projects

Even before the complexities with unreliable engines occurred, the risk of project failure was enormous.

It wasn’t funded sufficiently. Back then, there were multiple concurrent space programs, and landing a man on the Moon wasn’t even a top priority. All the funding had to be distributed. Contrary to that, the American Apollo program was a single, nationwide effort.

The social acceptability risks were not mitigated. The project’s sponsors (the Soviet party leaders) had lost interest in the whole idea after the United States landed first.

The technical risks weren’t mitigated. The idea was to do a series of test launches and fix any occurring problems. In software development, this approach is called the “big bang integration testing,” and is strongly not advised. The testing facilities for engines and an entire rocket weren’t built because the funds were limited. Due to the engine’s design, it could only be fired once and couldn’t be reused afterward. This made the rocket launch the first and only way to verify if the engine was working properly.

You can find all the risks I’ve mentioned in modern IoT projects. Surprisingly, the technical risks are not the main reason why IoT projects fail. The organization culture and ecosystem partnerships are crucial for any new technological endeavor.

Fifty years have passed since the race to the Moon, but the same principles of failure and success hold true till this day.

Want to learn more? 

If you are interested in AWS — check out my free AWS course on Udemy.

Acknowledgments

Author thanks Racoon Writing for copy-editing this material.