YTread Logo
YTread Logo

Why it Was Almost Impossible to Put a Computer in Space

May 30, 2024
Between radiation, microgravity, and extreme power and cooling limitations,


is the worst possible place to put a


. Can you let me in? That's better now, if you follow


research and exploration, you probably know that space is full. of


s like this, so how do they do it so well? The short answer is by replacing them many times. The ISS receives regular shipments of dozens of laptops at a time and this is great. This is a direct quote, absolutely destroyed, but not all computers. can be disposable and in 2017, the folks at Nasa HP Enterprise and kokia, who sponsored this video, collaborated to create spaceborn one, the first Edge Computing server that was intended to operate for an extended period of time on the International Space Station, for Of course, being the first.
why it was almost impossible to put a computer in space
I tried something, let's say a relearn took place and it turned out that the supercapacitors in the SSDs were prone to radiation related failures, who knew, but since all I had to do was run benchmarks anyway, the mission was terminated. considered a great success and in 2021 They released Spaceborn 2 whose purpose was to go beyond proof of concept and explore practical applications for early computing, especially data analysis using AI, but the story does not end there behind me. New Space Born 2, for administrative reasons, has the same name and core. specs like last time, but it took off earlier this year and has more storage than ever, over 130 tablets, which is an incredible feat considering the design challenges.
why it was almost impossible to put a computer in space

More Interesting Facts About,

why it was almost impossible to put a computer in space...

I mean, where do they install these things? Look, sure, I guess not. You don't really need a ladder, do you? But you could illustrate why Edge Computing is needed on the ISS. Let's look at a use case that focuses on astronaut safety. These are the Eva gloves that crew members wear during spacewalks and accordingly. In a 2016 article, they were responsible for half of all spacesuit injuries, so to ensure their integrity to each other, NASA requires the crew to take hundreds of photographs of them from all angles and then transmit them back to Earth, where machine learning is used to analyze them. for scratches or other hazards, except for a small problem that data transfer takes five days, but with space transportation, oh I don't know how, 45 seconds is not only a huge time, but only a handful need to be sent of photographs to Earth.
why it was almost impossible to put a computer in space
Further analysis Spaceborn can free up a significant amount of the crew's limited network bandwidth for other, more interesting things with such obvious benefits, so one might wonder why no one tried putting a server on the ISS before the short answer be after seeing. how the crew's laptops fared, a lot of people thought the plane just wouldn't work and even if they did, there were a number of other obstacles to overcome, like the launch, okay, this is really cool. Rocket companies like SpaceX and Northrop Grumman have Shak test machines that are programmed with profiles that will simulate the launch conditions of their respective rockets and if you've seen that viral video of the machine that dismantles hard drives by shaking them vigorously you'll know that surviving that type of treatment is no small thing.
why it was almost impossible to put a computer in space
Well, these machines managed IT both in the simulation and in the real world. In fact, they took it off at the end of January. Each piece of equipment must also pass an acoustic chamber test and usability evaluation in the US to ensure that station personnel can install and manage it and, oh, here's a good one, apparently all equipment shipped to the ISS goes through what is called a white glove test which, fortunately, is not what it seems - basically, you put on a pair of white gloves and then you just handle the trash. If your gloves catch or tear on something, yes, that is a potential source of injury.
I'm going to need you to register it, which they actually do on site and then repeat the test. I just hope they weren't filing. Speaking of any RAM, let's take a closer look at these machines, the Z machine, because it's not just one here. I know that was the goal of this whole experiment, but it still weirds me out that these are just standard HP Enterprise systems that you can order on their site today. They don't even have lead armor or anything. In this case, we are looking at an Edgeline 4000 which is a multiblade system and a dual socket dl360 server.
We asked why these specific machines and the answer. What we got was surprisingly identifiable, we sorted the HP server catalog by deep power consumption and GPU support and these were the ones we were left with, fair enough as to why two different machines, well, here's the thing in a world perfect, multiples of the same machine. It would have been better, but due to power limitations, they chose to have one with more CPU cores for more additional scientific applications and one with pure CPU cores but with a GPU for deep learning and artificial intelligence. However, one thing they both needed is generously ample Kokia storage. sponsored this and brought us here, so let's take a look at the, let's call them, unique choices they made for their storage configurations.
Firstly, obviously any trace of supercapacitors is gone, so Kokia can proudly say their SSDs are ready for space I guess, but what's less obvious is why they chose a SAS interface drive instead of nvme for your high speed mass storage. I mean, you'd think this was space age technology. They would want it as fast as possible, but these drives were selected for their balance of performance. reliability and especially power efficiency when looking at a power budget shared between two servers that is less than that of a typical gaming rig, every batch counts, of course, and that's even under ideal conditions at any given time to conserve power for other priorities At the station, the equipment may be asked to operate in half-power mode or even shut down completely for large operations such as docking, so the new DL 360 server, this type receives four PM6 Enterprise units from 3.72 TB totaling 120 tablets of raw mass storage for scientific data. and for backups, then for application drives, again we have a really wild setup.
They chose four units, but this time it's their RM six again. They are using SAS for lower power consumption, but this time two of the drives are running redundant data. mirror and the other two are basically cooling down there, ready to be sprung into action in the event of a failure, two hot spares from a set of four units would sound crazy Paran oia on Earth, but I assure you that in space where the bit changes random radiation is much more common, it's perfectly reasonable, I mean other than the excessive Drive settings and the 28V power conversion they need to run it on the ISS, there's not much to say about this, it's more or less a standard server, there is something cool. demo they said we could run, although oh yeah, they offered to let us take one of the drives out of this dummy machine and live swap it into this running machine to prove that no data loss will occur.
You want to do the honors, I'm sure everyone does. What they asked you to do is put it in Bay 8. Look, you can see that the unit is actually working. Use the Bay s that works. Look, I mean, that's good, rm6. That would be bad if it wasn't Kokia and then let's check the size. Beautiful 3.8. tab exactly what we want, it's not the exact units that are on the space station, but we wanted a different capability to show you that it's working and that they are expensive, look at that state, rebuild, I mean, this seems like a lot of extra steps that we must continue. you could have just looked up the light, yes the light, yes it's going to be a hit and whether you're looking for a SAS drive, a high capacity or high performance nvme drive we'll have plenty of Kyo drives that have great enterprise level drives linked in the description below I think I'll let you take this one apart, it has an appropriate line of this size, okay let's look at the second server that is packed into each locker, the el4,000 is a blade chassis, so servers are basically these little sliders. on cards that yes, I know AR, they are cute or what wait, take them out and they come in from the side, look at it instead of the front, wow, and they managed to fit four kokia xg6 nvme drives into each of these blades, well when I say Each of these blades, I should say, had the power budget for four drives, but they didn't have the power budget for four blades in the flight configuration of this system.
They ship with only one of the four blades installed, although it should be noted that I stand up one spare blade per system in the event of a failure, there is simply no way that poor Locker can support both these blades and the other server running simultaneously . Let's save it and shift our focus to the locker. Now obviously there is no real improvement. or downward, as it orbits the Earth at about 28,000 km per hour, but to improve the comfort of astronauts, they tend to mount directional elements, such as plants, in a fixed orientation that will position our lockers and the drawers that hold our servers in the roof.
There are two of these lockers, each containing an identical system load for workload sharing and redundancy, and these lockers present some serious design challenges, starting with the fact that they use a standard that Lally does not exist in. the Earth Express rack for packaging servers. Then HP Enterprise had to get a little creative, they found the shortest servers they could and then they put them on their sides and are using a combination of air cooling and water cooling. The air cooling uses a system on the ISS called AAA, so at the back of these lockers there are two cold air supplies and then two hot air returns that handle about 20% of the cooling for the servers, obviously 20%, not 100%, they're going to need a little more and that's where this water cooling comes into play. this is not one: one about how it would be deployed on the ISS, for one thing, these accessories, 3D printed mockups, are really worth $800 each, if you were even allowed to buy them, these cheap vinyl tubes from Home Depot, they are They must make real tubes. stainless steel, in fact any wet surface, so anything that comes into contact with water is supposed to be made of stainless steel, but we can still illustrate how it's supposed to work, so on this side they're going to a heat exchanger very similar to this one.
It is actually from the first space generation, but functionally it is the same: it pumps cold water here, cools the air inside the system and then takes out the warm water to dissipate it in the photovoltaic heat exchangers that are connected with liquid ammonia refrigerant and mounted on the outside of the station to sink that heat into space you need these kinds of special heat exchangers because even though we think space is cold and we see people you know ejected from the airlocks and freeze or whatever in movies, the truth is that for traditional heat dissipation methods you need air and in the near vacuum of space, well, there is no air, get it, corny jokes aside, the two cooling systems together are good for remove approximately 400 watts of heat from each locker, but that's a combined budget.
So if this GPU server speeds up, these CPUs better relax a little. Now let's talk about one of my favorite topics: networking, there are four standard RJ45 ports on the front. Okay, I'll show you live. one, as I was saying, four ports in each locker, two of them connect both internal ISS gigabit networks to a separate redundant switch inside the locker and then the other two links will go between the two lockers at 10 GB, why 10 gigs ? well, because for backups or multi-node workloads that's way better than gigabit and the power budget didn't allow for anything faster.
I think I'm starting to notice a pattern here anyway, it's all pretty standard but things get even less so when you look at the station side of these networking cables, this is a 37 pin mil spec locking connector They are designed for power and data, but in this application only eight of the pins would be used and now it costs $220 just for this part. On the space station, NASA provides these cables for you, but to test here on Earth, HP Enterprise had to make their own fantastic, what will really blow your mind is that, despite all their expensive networks, thesemachines do not have a normal Internet connection. just a private link back to Earth that NASA not only limits to 1 megabit per second, but they also encourage people not to take full advantage of, and even now in 2024 it doesn't have 247 connectivity pretty much every hour or two. a period of inactivity that can be as short as a few minutes or as long as 45 minutes and that is because they have to prioritize generating enough power for the station and when the giant solar panels are pointed towards the sun they can block the line Of vision. with satellites providing connectivity, which is a fun fact, even though the ISS orbits less than 500 km from the Earth's surface, our ping times to the ISS and yes, we did get to ping the ISS , which was pretty good, but our ping times were an atrocious reach of


a second when we uploaded some of the best memes Earth had to offer LT Now we asked why that is and the answer was twofold: one , it's very old, okay, but also the station's Internet relay is in geosynchronous orbit over 35,000 kilometers from the Earth's surface and, well, what I said was H, well, there's your problem, it's too far away and you might wonder, well, why don't they use Starlink?
That's a good question, they might do it someday, but. For now they don't and HP and the team on the ISS have to work out the limitations of the current setup. I mean, it took them four years to validate that we can run even a normal computer here. and they really expect this to be reliable, they can't just switch to something and leave, I don't know, I hope it works, um, oh, for The way here is another fun one: there is no API to determine if your connection is up or down, so which what they do is buffer all their communications in basic terms, meaning they ping every second and if the ping is successful they send data and then wait.
That the connection was maintained during that time is a pretty good system, well it's not perfect but it's certainly enough for us to learn a lot from the spaceborn project even though spaceborn 2 has been in action for 3 years there's still a lot to do. learn two of the servers have hard hardware raid cards for their drives, for example, that cost mass and power consumption, while two of them use software raids that obviously don't consume mass, but could affect power consumption even more. power depending on the loads and one of them could be more or less reliable than the other, we won't know until we try, which is a recurring theme here, so if you want to learn more about the space computer project, We'll have some resources linked below and we'll also have an L for some great Enterprise storage options from our sponsor Kokia.
We're really grateful for this unique opportunity to realistically get as close as possible to the real ISS, a prop on a soundstage in Los Angeles, but oh well. Thank you Kogia for the opportunity and for your long-term partnership.

If you have any copyright issue, please Contact