Vulnerabilities in space.

[MUSIC] A bad update to CrowdStrike security software has caused massive systems outages around the world today in just about every business sector that you might imagine. Right now, we don't have details on how this outage has affected space systems specifically, but no doubt we'll be hearing more in the coming days. In the meantime, if you'd like to learn more about this incident in depth, I encourage you to download today's episode of our sister podcast, The Cyber Wire. >> You know, I did wonder this morning, Maria. I kept knocking the space bar and I'm still here on Earth. [LAUGH] >> Nice. [LAUGH] [MUSIC] >> Today is July 19th, 2024. I'm Maria Varmazes. I'm Alice Karouf, and this is T-minus. [MUSIC] >> The National Security Space Association releases a GPS vulnerability report. IM2 picks a lunar landing spot. Rocket Lab delays its next electron launch. >> And our guest today is Yanni Bagouti, CEO and co-founder of Cosmic Shield Corporation. Yanni will be talking to Maria about providing advanced shielding technology to protect graphic cards in space. It's a fascinating chat, so stick around for the second part of today's show. [MUSIC] >> Happy Friday. [LAUGH] Let's get into today's headline, shall we? The National Security Space Association's independent think tank, the Mormon Center for Space Studies, has released a report called America's Asymmetric Vulnerability to Navigation Warfare, Leadership and Strategic Direction Needed to Mitigate Significant Threats. The report is authored by Mark Berkowitz, a member of NSSA's Board of Advisors, and describes growing threats to the US global positioning system. It also goes further by making recommendations to mitigate vulnerabilities to such threats and reestablish US leadership in space-based and terrestrial positioning, navigation and timing, also referred to as PNT. The report starts out by explaining that GPS is a global utility integral to economic development and growth, transportation safety, critical infrastructures, and US and international security. Unfortunately, GPS has been surpassed as the premier space-based PNT system in the world and is vulnerable to a variety of threats. It warns that further, the United States does not have a reliable and resilient terrestrial backup of GPS, while other nations, namely China and Russia, do have backups for PNT services. Now we're not going to go too deep into the report, but we have linked to it in our show notes and strongly suggest you check it out for yourself. >> Can you just imagine if the GPS system went down a mirror to today's crowd-strike issues? There are whole generations like mid-millennials onwards who haven't learned how to read maps and don't even get us started on how we'd ensure our prime day deliveries would arrive on time. >> I am sure the thing pieces are being written already. >> Yeah, anyway, moving on, Intuitive Machines, which was the first commercial company to soft land on the moon earlier this year, has finalised the IM2 mission landing region ahead of its sold-out second mission. The IM2 mission is designed to prospect for water ice and other volatiles on the moon's south pole, which requires landing on a site that supports the high probability of ice stability within one metre of the lunar surface. Working with NASA, Intuitive Machines selected a 200-metre diameter elliptical region on the Shackleton Connecting Ridge with favourable terrain, earth communication position and solar angles for power generation. To align with the landing site's solar power conditions, the mission must be timed between November 2024 and January 2025. IM2 is currently planned for late 2024. The Polaris Dawn crew are patiently waiting to see how the investigations into the recent Falcon 9 mishap will affect their launch schedule. The commercial crew recently completed a series of space suit acceptance tests in preparation for the mission's extravehicular activity, also known as a spacewalk. The crew plans to complete the first commercial extravehicular activity in low Earth orbit during their mission, which was scheduled to lift off later this month. And speaking of launch delays, Rocket Lab's next electron launch will move to a later date at the request of mission partner Capella Space. The press release announcing the rescheduling insinuated that Capella Space needed time to complete additional testing for their mission. Rocket Lab isn't going to let the change affect their momentum. They say they'll be moving on to their next mission in the manifest with Sinspective, now flying next on Electron within the next few weeks. Momentus has entered into a convertible note and loan agreement to support operations and pursuit of funding opportunities for the company. The US-based commercial space company and space infrastructure ventures, which is a firm that invests in disruptive high-tech space tech ventures, entered into a convertible note under which Momentus may borrow up to $2.3 million prior to September 1, 2024 into tronches subject to certain conditions. Six directors and officers of the company also recently entered into agreements to loan Momentus an aggregate of $500,000. There's some good news for L3 Harris shareholders. The board of directors of L3 Harris technologies have declared a quarterly cash dividend of $1.16 per common share payable on September 20, 2024 to shareholders of record as of the close of business on September 6 of this year. Yay for payouts. When one door closes, another always opens and that seems to be the case for SpaceX employees that are not wishing to make the move to Texas. SpaceX announced that it would be relocating its California-based operations to the Lone Star State earlier this week, and other space companies are scrambling to offer employees looking to stay in California a place for them. AstroForge, for example, have offered openings in the same area as SpaceX's Hawthorne headquarters, and there's even been offers from companies in other countries, such as Latitude, which is based near Paris. Paris or Texas? Paris or Texas, but maybe not Paris, Texas. Just to be clear. Now, NASA and Boeing have provided a very underwhelming update on the Starliner yesterday. Hidden in their blogs was the announcement that the US Space Agency and Prime contractor have completed the Starliner engine test at NASA's White Sands Test Facility in New Mexico. Just down the road for me. Ground teams fired the engine through similar flight conditions the spacecraft experienced on the way up to the space station. The ground tests also included stress case firings and replicator conditions Starliner's thrusters will experience from undocking to deorbit burn, where the thrusters will fire to slow Starliner's speed to bring it out of orbit for landing in the southwestern US. The ongoing ground analysis is expected to continue throughout the week. The team is continuing to work through plans to return the Starliner from the ISS in the coming weeks. NASA is looking to expand the Space ROS repository with new higher fidelity demonstration environments and additional capabilities. Space ROS is an open source software framework derived from ROS2, which was created to be compatible with the demands of safety critical space robotics applications. And you will find the US Space Agency's challenge details by following the link in our show notes. And speaking of the selected reading section of our show notes, you'll also find two additional pieces that we've included today. Both are focused on the US Space Force and their response to threats in space. Hey T-minus crew, tune in tomorrow for T-minus Deep Space, which is our show for extended interviews, special editions and deep dives with some of the most influential professionals in the space industry. And tomorrow it is International Moon Day. We'll be sharing my full chat with Amanda Lee Falkenberg, talking about the Moon's symphony. Now check it out while you're out enjoying the weather, commemorating the date of the first moonwalk, or rebooting your Microsoft machines at least 15 times. You don't want to miss it. Keep hitting that space ball. Our guest today is Yanni Bargoudi, CEO and co-founder of Cosmic Shield Corporation. And I started by asking Yanni what motivated him to start his company, focused on shielding GPUs in space. Our co-founder and CTO, Dr. Lumbett-Seaver, he's been basically leading the field of space radiation physics, dosimetry, target fragmentation, basically how it interacts with the human body, materials and so forth. And we met a few years ago in the summer, late summer of 2020, talking about how essentially space was exploding in the design industry. And it was really under a lot of people's noses, right? It wasn't very popularized at the time, just how many more spacecraft were going up into orbit than ever before in history. And of course, it's increased since then. And we were talking about how essentially the way the spacecraft are manufactured, the materials used, some of the fundamental materials, are pretty much the same as they have been for decades. At the same time, everyone's trying to do crazy new things with these same structures. They're trying to make on-orbit servicing, advanced robotics, space stations where people can live for decades on orbit, and nuclear propulsion even, plasma engines on, new spacecraft designed to go into deep space. So we were saying that it's what we think a significant inefficence. And the problem with space radiation is that for a long time, well, one is a very niche field, right? There's not many physicists who specialize in target fragmentation of space radiation. What you need to understand, you know, most of the shielding mechanics, right? I can get a very cursory overview there. And in addition to that, there was a lot of focus on the more flashy things, rockets, you know, the big engines and the giant space stations that left a lot of the infrastructure funding kind of in the dust. Not quite. It picked up a lot over the past two years. So that's very big positive. But what that led to is essentially a lot of traditions being followed with spacecraft design, right? The idea that you can't really shield to the point where you can have an NVIDIA GPU last the same amount of time that a $250,000 RAD hardship would last in space. Well, in reality, though, the physics has made quite a bit of headway. And we have seen that not only is it possible, it is feasible to do so in a way that is easy to integrate because you can have the best in the world. But if it's hard to use, you know, like it's going to make your spacecraft 10 tons heavier, right? Or it's going to like you might have heard about using water and hydrogen as a way to shield astronauts, right? How do you engineer water and hydrogen in tanks around a spacecraft? That's a little bit of a headache for engineers, right? We would make something easy to use. We worked with an MIT Polymer and Metabterals group called the Boiski lab for some time to do some of the fundamentals on benchmarking, being able to test some of our ideas that allowed us to move pretty quickly. We were doing quite a bit of benchmarking because again, this is a really old field that's being disruptive. And naturally, the bar for proof is very, very high because it's something that people have been used to for decades, right? So, yeah, long story short, we were, we did a lot of accelerator testing at particle accelerator facilities across the world, saw very favorable results. And then starting of last August, honestly, we started commercially producing. We had a few launches. We had one with Axiom in May 24th last year. That was a demo, eight months. We returned it in March. We sent another system up in March and with the quantum space. And on the SpaceX transporter 11, we actually had scheduled to launch the first, you know, Nvidia jets and, you know, 100 teraflops chips of its kind into orbit. So my goodness. Wow. OK, thank you for that great overview. A bunch of questions came up. I'd love to get into all of that. So I just want to get a little more context for myself, really, because we're talking about, you know, AI, GPUs, essentially going into space, enabling a spacecraft to be more autonomous, doing incredible processing on orbit and beyond. Can we just sort of set up like the context for or the challenges different in shielding something like that? Then, you know, can you get into what those differences are? Definitely, definitely. So the we'll start from the radiation environment and space, right? A lot of people think of x-rays, gamma rays when they think of radiation, you know, from like a nuclear reactor or when you go to a doctor's office, for example, in space, the main concerns are particle sources, right? So you're talking about protons from the sun, solar protons, protons that get trapped from the sun into the earth. Micronutosphere, cosmic rays, you know, which are basically particles, massive particles, massive in a physics sense, right? Yes. I'm going to launch that near the speed of light, you know, coming from supernova across the galaxy, right? These particles are a lot harder to stop than a lot of the EM sources we see for radiation. And the problem with these particles is when they hit a target, whether it's a DNA strand or a transistor or even a pixel on a camera's CCD, right? They can either cause rapid change in the voltage on the device that causes it to trigger a different reading. So turn a zero to a one and vice versa that can cause an error. They call those like soft errors or it can actually destroy the target, which is a destructive event effect, right? So the problem with GPUs in general, right? Most radiation hardened electronics that we've been using, funny enough, they rely on old architectures. So a larger transistor is the further spaced apart it is. The it gives you some reliability against radiation. The smaller gets the opposite effect. So Moore's law is not good for space. Right. OK. All right. Interesting. It is. It is. Yeah. And other ways they get around this is the error correction codes. Sometimes they catch an error, they rerun the calculation three times, make sure each result is the same. Sometimes they physically triple up the number of bits on the processor and run in parallel to make sure everyone's getting the same result. Sometimes they send the signal three times. Right. There's a variety of techniques, but at its sound like all of this reduces throughput by quite a bit. So with GPUs, graphics cards, AI capable systems, you know, they're parallel architectures. So an error in one place has a much higher probability of causing cascading errors across the chip. Let's say the chip to simplify. The computer engineer guys probably look at this and be like, use better language. It's fine. I think for our purposes, we're OK. Yeah. Yeah. Right. So for that one reason, simplification here, it makes them particularly vulnerable to upsets. So the risk of upsets can cause much higher probability of causing catastrophic failures or a hard reset, for example. And then, of course, a lot of these new chips are using cutting edge like process nodes and really small transistor sizes are very densely packed. So then that's a physical issue too that causes probability of a heavy ion or proton having enough energy to trigger destructive event or just a soft error on that part is high. So the shielding requirements become much more severe, almost like you're dealing with a human being, because the same physics applied to DNA. Right. It honestly applies to most vulnerable components in space, including biological components. Wow. Yeah. OK. Exactly. And that's a simplification. Basically, they have high vulnerabilities for those reasons. You sort of hinted at this, that maybe material science hadn't caught up to that challenge yet. So you all had to, I imagine, sort of create something that maybe didn't exist or a solution that hadn't happened yet. I mean, how do you arrive there? I mean, that's amazing. So the physics, you know, were in the physics world to understand detailed mechanics, like these things called quantum molecular dynamic models, QMD models with Monte Carlo codes that basically give you good statistics to get very accurate simulations on how these particles interact with matter and how they produce secondary effects and all of this. That's what Lumbett, our CTO, that's what he's famous. A lot of fundamental physics models there. Now, in a physics sense, they're relatively new, you know, that degree, that fidelity of understanding of these reactions. But, you know, we're talking, we're still talking physics. So it's like 20, 25 years we've known about this. And there's always a lag between physics and engineering. Right. So the interesting thing is a lot of the fundamentals of what we've done has been studied in the past. So a lot of our team members actually worked with NASA, Dragon Space Radiation Analysis, you've heard in the past. I've worked with HESA and JAXA on radiation mitigation strategies and techniques. But the commercial application just had a bit of a lack. And that's partly due to the fact that one, the sheer cost of satellites was at a point where it was kind of not the biggest concern. You know, we were talking about, you know, in the early 2000s, mainly satellites were these massive bus-sized, like school bus-sized vehicles where they can pack in a ton of shielding, use these, you know, $40 million, $1 million hardened chips. They don't care. The whole system is $300 million. Right? Yeah. Yeah. However, with the advent of, you know, cheaper access to space, people not only wanted to send up, you know, more advanced systems, they wanted to send up more of them, right? And then at the same time, you get a lot of crowding in space. So that means they want them to last longer too, while being cheaper to make the ways. So these things created like a perfect storm for the physics to be applied. So we didn't do anything magic. We just applied the cutting edge target fragmentation, radiation, physics to the materials, created this nanocomposite polymer that has, it's a proprietary polymer blend that has a variety of nanoparticles evenly dispersed throughout it. One of the big tricks was actually we 3D print the material. We also, you know, injection mold it. But after you do that, how do you make sure that nanoparticle matrix stays somewhat uniform? That was quite a headache. Honestly, that was where most of our time was. Yep. Okay. Yep. We figured that out. We benchmarked it at accelerated facilities and we saw that compared to the next best they like even aluminum, which is the commonly used shielding material. They just make thicker solid aluminum, right? It blows it out of the water. The dust savings, 30 to 60 percent for the same mass. If you want to optimize for mass, you can get significant weight savings. But the big thing which people initially didn't see coming is that there was this misnomer that shielding is only good for dose, not for reducing the number of these protons, the damage potential of these cosmic rays. And that is based on old ray tracing models, old simulations, experiments only done with metals that were used to create a benchmark for other materials. Because the thing that's very important to consider is metals actually make it worse. Right. And the heavier the metal, the worse the field can get. And this is because when these particles fly in and hit your target, they can actually cause fragmentation of the atom. They have that much energy that they can cause neutrons and protons to fly off. Right. And in that process, you create gamma rays, you create X-rays. This is target fragmentation. That's what that is. So mitigating that is key. You need a really lightweight material. That's where the idea of using water or hydrogen even came from. Right. But then again, you know, how do you pack all that hydrogen in, which is effective? That was the trick. So I know I'm going on a lot of tangents. Like there's a lot of. Okay. I'm absorbing it. I'm honestly really fascinating because you're talking about a whole new paradigm here. So I'm just like that. That's got to be a fascinating, as you said, not just a test, but also the application of that now. And you all have applied it several times. It's flight tested. So I'm just like, that's, that's fascinating. So tell me about that. Yeah. So one of the big things we want to do with our flight test is not just corroborate the stuff we did on the ground. So all these particle accelerated tests, but we also wanted to showcase, look in space in a non imperfect environment, right? Because a particle accelerator, you can control a lot, right? In an imperfect environment, you'll see the same trends. You'll see the dose being reduced and you'll see the types of particles behind the shoe because they're, think about dose, like when people say radiation dose, like, you know, C, Vir, Rad, and so forth. Right. The dose is literally just an average of energy in an area. And a very important thing to consider is that you can have a high dose with a low damage potential that actually is possible. So think about it like tennis balls hitting a wall. If you have a contract, right? And then a bullet, the same kinetic energy, but what's going to go through the wall? The ball, same company. So we wanted to measure every single facet. So we were not only looking at the dose, we were looking at the LET, that's the amount of part of energy, the particles that actually make it through the shield have when they hit the target, right? The amount of energy they transfer. The overall average energy, the distribution of the energies of the particles going through, the number of the particles going through and see what effect that would have. And what we saw is that in a lot of cases, the number of protons that actually make it through is actually reduced by a factor of 10. The energy potential, let's say, the LET, it shifts to the left on the chart, which means like, you know, you are actually the protons that do make it through are less damaged. If you just looked at dose, you might actually miss this, right? And if you just use the metal, funny enough to benchmark your shielding models, you might have missed this. So then if you said, okay, we use the metal because it's producing a lot of secondaries, which again, have a high LET, right? High damage potential. So if you put a chip behind those metal shields, you would see potentially a reduction in dose, but you would see no change really in the air. Right. However, if you were thorough and you conducted this test dozens of times over with a variety of materials, many controls, you would see this effect. So that's what we did. Right. And we actually not only with the ones on orbit, we have 48 systems similar to the ones that's launching the NVIDIA system that's launching on the transport 11 mission that are going to be tested with the Air Force over the next few months, just to get an even better characterization of what's going to be happening on. We'll be right back. Welcome back. And what with the massive outage today, it feels like a not so great day to be encouraging anyone to do some hacking, but the world's largest global hackathon has opened its registration as of today. It is the NASA space apps challenge 2024, a hackathon for coders, scientists, designers, storytellers, makers, technologists and innovators around the world to come together and use open data from NASA and its space agency partners to create solutions to challenges we face on earth and in space. This NASA hackathon is hacking as in hacking a solution together quickly and not hacking as in taking down global systems, just to be clear. Last year, the space apps had nearly 300,000 registrants in over 185 countries and territories and worked with 13 global space agency partners. It's a two day hackathon event this year happening on October 5th and 6th, which is a weekend, by the way. This year's theme is the sun touches everything and participants will be working with NASA heliophysics to take on some interesting challenges. And I should note this is in all ages, all skill levels and all backgrounds challenge and it is free to try. So really, if you think taking part in a space themed hackathon sounds like a fun way to spend a weekend with friends and stretch your skills. And honestly, yes, yes, it does. Registration is open right now. [Music] Well, that's it for T-minus for July the 19th, 2024, brought to you by N2K Cyberwire. For additional resources from today's report, check out our show notes at space.n2k.com. We really want to know what you think of our podcast, your feedback and shows. We deliver the insights that keep you a step ahead in the rapidly changing space industry. If you like our show, please hope you do share a rating and review in your podcast app. Please also fill out the survey in the show notes or send an email to space@n2k.com. We are privileged that N2K Cyberwire is part of the daily routine of the most influential leaders and operators in the public and private sector. From the Fortune 500 to many of the world's free eminent intelligence and law enforcement agencies. N2K makes it easy for companies to optimize your biggest investment, your people. We make you smarter about your teams while making your team smarter. Learn how at N2K.com. This episode was produced by Alice Caruth. Our associate producer is Liz Stokes. We're mixed by Elliot Peltzman and Trey Hester with original music by Elliot Peltzman. Our executive producer is Jennifer Iben. Our executive editor is Brandon Karth. Simone Petrella is our president. Peter Kilpey is our publisher. And I am your host, Maria Varmasas. Thank you for listening. Have a great weekend. . God. . . (whooshing) [BLANK_AUDIO]