Well, binarily speaking of course! The 4th of October 2021 marked 64 years since the launch of Sputnik-1, the first artificial Earth satellite. The 84kg spacecraft — launched by the USSR on board an R-7 rocket — worked for roughly 2 weeks, and it reentered the atmosphere a few months after. Its architecture was, at the most, rudimentary: its batteries weighed 51kg, it was equipped with a 1-Watt transmitter which encoded telemetry in low-frequency pulses — which would eventually be broadcast and heard on AM radio — and it was pressurized with nitrogen. It’s safe to assume that Sputnik’s architecture was exclusively crafted for this history-making project. It was surely worth coming up with ad-hoc contents and subsystems for it.
Sputnik was a massive technological achievement and it paved the way for an industry which changed forever the way we communicate with each other and the way we observe our planet, ourselves, and our Universe. 64 years since Sputnik, nearly 10000 satellites have been launched, of which 3000 are currently alive. Most of them don’t really look like Sputnik (space designers quickly abandoned the shiny spherical form factor for more complicated squared geometries). They are surely more sophisticated. Not only in terms of the obvious more generous computing resources and data rates — which have skyrocketed throughout the years — but also in the way they communicate to their home bases. They don’t, in most cases, unsystematically broadcast radio energy anymore, but they communicate by means of point-to-point encrypted links from space to ground.
Sputnik-1 was as much a pioneer as it was a loner: it floated alone in space whispering radio pulses until it perished out of exhaustion (its silver-zinc batteries died).
Are modern satellites that different compared to Sputnik-1? The answer -perhaps surprising- is: not really.
Satellites are still loners. They are largely unaware of the existence of other satellites, let alone being capable of sharing data with each other. Space systems are not interoperable, and rather uncomfortably disconnected. Imagine how the Internet would have turned out if computers were kept isolated from each other. That’s space today. Space hasn’t yet experimented much with the power that networking brings: when we interconnect systems, the sum is greater than the individual contributions of the parts; the physical and functional architectures depart from each other, and functionality becomes a distributed factor. The individual emerges from the collective. Space is realizing this, standing at the verge of a tipping point. Want evidence? See Elon Musk’s tweet from some weeks ago when he explains Starlink satellites will start to talk to each other:
Starlink will soon become an in-orbit mesh network. What’s more, Starlink will become a massive spatially distributed data router. Still, Starlink satellites will talk mostly to other Starlink satellites. The true change will come the day an Earth Observation constellation operator will be able to launch a satellite and make it connect to Starlink or any other networking provider and make the data flow seamlessly. Or when Starlink network will be able to route data through any other spaceborne network. It’s about time we just let data flow in-orbit. It seems to becoming.
While Sputnik-1 itself as a milestone should be celebrated, Sputnik-1 as a paradigm of a lonely and disconnected space must serve as a compass of what needs to change. With activities such as asteroid mining, the Moon economy or setting a foot in Mars getting under the spotlight, there is no way we will get there by launching disconnected loners.
What are we still missing? How do we bring satellites together in a more cost-effective way?
We need a combo: modularity, autonomy, connectivity, and availability. Let us explain.
We said Sputnik-1 internals surely was ad-hoc, and rightly so. Today, spacecraft internals — for almost every mission — still are ad-hoc, 64 years later… Spacecraft components remain a very specialized industry. An expensive and slow industry. Spacecraft and their components are literally handcrafted, they are exactly the opposite of what a commodity is. This makes space projects dull and captive of specific suppliers. But this will change in the near future. Spacecraft will be built similarly to how gaming PCs are built today: you have a budget, you need components, there is a marketplace, you select the components you need according to the budget and needs, you assemble the components, you install the software, off you go. Too far-fetched? Not really, see how companies are already establishing high-volume production of spacecraft components. Commoditization of space components is a tangible reality and nothing indicates the handcrafted way will stay for long.
Modularity has withstood a recurrent critique from space systems designers throughout the years: it’s heavier to be modular. It is usually said that modular architectures have comparatively higher mass, and that might be fairly true: when you use a generic approach, the penalty you pay by going generic is that your solution is not exactly optimized for a particular scenario: many miss the fact that this is precisely the beauty of it because it provides the freedom to use the same common building blocks for a variety of situations.
The mass has been historically a challenge in space, simply because slinging a gram of mass into orbit costs money. No surprise there. Now, the cost per kg of mass has dropped in the last decades, and nothing indicates this trend will change anytime soon. With more launchers making their way all around the world, quickly learning their lessons, and eventually reaching orbit, competition will be ampler, and the concern about streamlining mass will be increasingly a thing of the past. Or at least not such a salient design driver as it used to be decades ago. Just like Youtube became a feasible thing only when internet speed made video streaming possible, modular spacecraft will become a de facto thing very soon, when the launch market will offer more options and the price/kg will make the discussion about the penalty of comparatively slightly-heavier modular avionics a vintage one. When we talked about the gaming PC analogy, we mentioned the need for software at the end. As modularity will accelerate the commoditization of space components, it will also accelerate the need for more software to glue all the now commoditized components together seamlessly. Thus, modular architectures will necessarily be software-intensive architectures. Standardization will also be required as components will adopt common interfaces: commands and telemetry formats, data sizes, etc. The software will play a key role in creating the right abstraction layers to make sure standards will not become a blocking factor when it comes to interoperability among component providers.
Modularity can be even taken a bit further from just what’s under the hood of a satellite. Functionality from a space system has been historically bound to single flying boxes. This is in part because we are used to that. This is called functional fixedness: we keep on using things only in the way it’s traditionally used. Had Sputnik-1 been from the outset a flock of cooperative satellites, the space industry would look very different today. Space technology is evolving to make this a reality: what if functionality can be broken down into multiple elements flying in a vicinity? What if spacecraft function was not concentrated in one single physical body but spatially distributed and connected wirelessly by intersatellite links? In this configuration, different functional blocks flying next to each other couple their functionalities by means of high-speed wireless links. Then, a mission does not map to one satellite anymore, but to a group of them. And thus the risk gets divided as well. Degradation of any of those functional elements can be handled by selectively deorbiting them and sending new segments to the formation group, extending its aggregated lifetime. For an architecture like this to be usable, automated flight control and precise absolute and relative position determination is a must. But also autonomy, since a ground operator cannot babysit each element any more (imagine having a ten-minutes pass to resolve faults of a flock of tens of satellites). This yields the floor for the next elements of the combo: autonomy, availability, and connectivity.
Autonomy, Availability & Connectivity
There is an unspoken and undeniable symbiosis between spacecraft and us humans. Satellites have always taken us for granted. For 64 years, satellites have been machines operated mostly manually. Since we do not have continuous access to command and control spacecraft but only discrete-time windows, we equip spacecraft with the minimal autonomy to barely stay alive during the times we can’t really talk to them. This approach has worked fine for the “disconnected” space industry paradigm we spoke about a while ago. But it is definitely not fine for the connected, data-driven, and distributed future that is coming ahead. We will not be able to mine (or geez, divert) asteroids, or effectively colonize other planets if all the parts of the puzzle will require a human brain behind the wheel: we will need refueling stations, data relays, in-orbit manufacturing, landers, rovers, helicopters. All these machines will have to autonomously cooperate between them and with us if we want to realistically explore the space beyond our noses in a cost-effective way. We cannot have 24/7 teams of humans babysitting space infrastructure. Space infrastructure needs to work also while we sleep.
We live in such a protected bubble in the Universe that it makes us believe that space, at large, is all like this benign environment we live in at the surface of the Earth: the Magnetosphere simply does not get enough credit (except for aurorae). The Universe around us is a collection of particles of various energies. And some of those particles are so energetic that they can affect the artificial systems we send to space. This has been a known concern in the space industry since Sputnik-1 and before. Particles will interact with the electronics in orbit, and this may affect our space missions, to the extent of completely ruining them. And the solution is not just about duplicating or triplicating all the subsystems inside a satellite to get rid of the problem: increasing the number of components increases complexity, and switching between redundant elements (either hot or cold) does not happen magically; it takes time, and it can disrupt nominal operations and data flow. The future of space calls for high-availability systems which can maximize QoS (quality-of-service) autonomously in the presence of glitches and faults, which will continue happening. In the past, spacecraft have been equipped with failure-detection capabilities indeed, but availability (as in, uptime, or time in the nominal, revenue-making state) has been historically dependent on humans making sense of on-board faults, connecting the dots, and taking decisions to correct them, which can be slow and costly. With the advancement of machine learning and artificial intelligence technology, it is feasible to think that satellites will soon have the means to process their own health-data and estimate their current status but also to predict future trends, including self-diagnosing faults, and act accordingly to avoid them.
Goes almost without saying, space needs to connect. To interoperate. Today, space systems are like computers on Earth were during the sixties and the pre-Internet era. No way we’ll realistically reach out to other planets with disconnected systems. There is plenty of work to do in terms of networking and security protocols and standards, in-orbit data storage, virtualization, and a long et cetera.
Modularity, commoditization of space components, software, risk, and how they all relate.
Space as an industry is often held back by the principle of flight heritage - — ‘if it’s flown before and survived, let’s stick with it’. While it does make things safer, it has stagnated innovation after the initial frenzy in the 60’s and 70’s. It’s been a hectic 64 years of learning and innovation since that shiny 84kg sphere made it to orbit. We have engineered a gateway to space, but we still need to build strong foundations to make efficient and sustained use of it. We must look ahead to the next 64 years and start creating more modular, autonomous and connected orbital technology if we want our space exploration plans to become a tangible reality.
ReOrbit is founded on the pillars of Modularity, Autonomy, Availability and Connectivity, and has set out to lead space technology and usage into the next age. We are dedicated to the path of transforming technology and development to enable space applications of our current and future society. We adopt an approach that transforms the traditional thought of rigid and single-use satellites into reusable, flexible, distributed and low-cost space systems. Therefore, we develop technologies to make spacecraft platforms modular and configurable to allow fast time-to-market and accommodation of different payload types. Through a software-defined architecture, we unlock new functionalities such as built-in autonomous orbital capabilities.
ReOrbit — Space, Redefined.