Blake Moya
At the intersection of ideas and action, this is Longitude Sound Bytes, where we bring innovative insights from around the world directly to you.

I am Blake Moya, Longitude fellow from The University of Texas at Austin. In today’s episode we will be featuring highlights from a conversation I led with Minsik Lee, a Senior Technical Staff Member at IBM.

He is part of our Edge of Space series, where we explored the roles of individuals, experiences, and cutting-edge technologies that are preparing us for exploration on the Moon and beyond.

As a graduate statistician, I was interested to hear about the technical challenges involved in edge computing and the ways that overcoming these challenges can change the way we look at computing in general. We started our conversation with Minsik’s description of a CubeSat, a kind of miniature satellite on which his team at IBM would be launching their software.

.

Minsik Lee
I’m a senior technical staff member in IBM. I work as an edge computing architect, and I also work as a full stack software engineer. My focus area is mostly edge computing and hybrid cloud. For the last two years, I work in space tech industry. We are trying to apply our cloud native and the software-defined and emerging technology on the space industry. That includes the Space Station, or satellites, and this CubeSat.

Blake
Would you mind breaking down what CubeSat is and what this Endurance CubeSat mission is for people who maybe haven’t heard of it?

Minsik
The new concept is that instead of some expensive satellite, we can maybe launch some small microsatellites in the space and it can have an economic benefit. We can use a shared environment. There are many companies, including some universities, that are now developing and sending these Cubesats to space. It is a mission similar to satellite. The mission can be different based on the kind of equipment that is attached to the CubeSat. For example, some CubeSats can have some kind of a camera that can provide images of Earth. They also can have sensors. They can provide the location or temperature or some other sensor data, they can collect it there and send it back to ground. That way, we can collect sensor data in a space environment and then do some more analysis testing on the ground.

Blake
Thank you. And so what is your role specifically on this project or within Endurance?

Minsik
Yes. So our team focuses on edge computing areas. You may be aware that there was a project called Mayflower. That project developed an autonomous boat. From the IBM side, we applied IT stack and then also provided operation to support that autonomous boat. In our perspective, that kind of use case for the edge computing environment. And then one day we just thought, why don’t we apply the same technology onto space, so we just had that idea about two years ago. And then we met people in NASA, and then we shared these ideas. And then we have some opportunities to apply software defined technologies on Space Station or satellites. Then we started a project called Endurance. That’s our own CubeSat. And so we will launch this one next year. And then our goal is to have kind of a playground. That way anybody can test their code in our CubeSat environment. Of course, CubeSat is very small. We plan to have a kind of emulated environment on the ground, so they can test their code on the ground environment. We will also have historical data gathered from the CubeSat so they can test and do some AI processing with their sensor data. And then we try to open this kind of playground with this environment to anybody, not only for some special scientist. And then after the launch we plan to share our architecture and then our lessons learned in public.

Blake
Yeah, I love the motive of this Endurance misison of, I believe the tagline is “democratizing access to space.” And you had made a remark where you said, you know, taking the edge computing principles used in Mayflower and just bring it to space. And of course, the “just” there does a lot of work because getting to space is difficult, but this is a big effort to bring space to the masses. Do you want to go into a little detail about how people could- I know that you mentioned that it’ll be a playground where programmers and developers can work on their code, practice their code, and then send it up to the satellite itself? Could you talk a little bit about what kind of people you’re hoping to attract to use the CubeSat?

Minsik
For now, our primary target will be high school students, and then maybe students in the university who can help. Maybe their skill set will be different, right? Some people will have a sciences background or a teaching background. Some of them will have some deep understanding or a high-level scaling AI. So originally we are developing that kind of class and for the high school and the student first. And then we will provide a different level of languages. So they can just develop their code using Python or node.js. We will provide API and then testing development environment. The people who are not so familiar with coding, we will provide kind of a web-based, block-based coding experience, similar to Scratch. So we will provide that interface. Also, of course we will provide some kind of a guideline, and then some kind of step-by-step kind of a thing. So that way, they can imagine something, and then they can understand that they can focus on the problem. That way, they can just consume our data on the ground, and just build the code, and then we will have a system that will convert the block to Python code. Then we can send this code to the CubeSat. We will just execute that, and that we can get the result really quickly. However, sometimes it depends on the schedule of our ground station connectivity, sometimes maybe within a day we can get the result for that. The main thing is that we are trying to provide some computing runtime environment. That’s the kind of playground. Yeah, so we are applying some cloud native technologies, such as a container technology, Kubernetes technology, on our environment.

Blake
Yeah, I was reading about how this kind of code that people produce would be containerized to run on the satellite itself. And so if my understanding of edge computing is correct, it’s that we have this distributed process, there’s some analysis that we want to do on data here on Earth, there’s some data that we want to collect off in space. And we’re deciding where each computation takes place. And I assume the goal with edge computing is putting as much computation as we can up at the data collection level to limit what we have to transmit to and from, is that correct?

Minsik
Yes, correct. And then on the ground we have more computing power on the system and backend. So you can do the kind of distributed parallel processing on the backend side with historical data. And then in space we are trying to find some use case that can have a benefit in the CubeSat environment. Sensor data, image data processing. And we are trying to include some other some interesting technologies, like quantum technologies, that way we can do different use cases.

Blake
And so then, a lot of your role has been in working in making sure that the containerization process, people’s code works and will run once it gets up into space.

Minsik
Right. I mean, in our network connectivity, we are able to only contact our CubeSat from the ground station. Each ground station will have 10 minutes’ connectivity to the satellite each time. So we are not able to send the entire container code to the CubeSat. We will have a pre-defined container runtime in CubeSat. And then from ground, we will only send the code. I need to make sure that we can execute on that code on the CubeSat in that kind of autonomous environment. If there is a crash or something, I mean, we are not able to use that anymore. So I’m testing. I need to make sure that nothing can break the operating system or runtime environment. Container runtime will provide limited access, and then the code should not break our CubeSat. So we think we need to have a kind of separate security policy, and then the runtime environment should provide just playground for the sandbox environment, for the code itself. Actually, that’s my main objective. Once we complete this one and we have a stable CubeSat environment, the next one is the ground-based development environment. That way, people have their class, their session using our ground environment. So now we are gathering some volunteers, developing a course and then also we are developing the ground-based environment functionalities.

Blake
And I mean, satellites seem like, to me, the pinnacle of edge computing, because it really seems like the highest risk of failure, because if it crashes up there, you can’t just go get it, you can’t go turn it off and back on again. Do you find it stressful, how confident you have to be in these processes in this software stack?

Minsik
Yes. We continue to test to make sure that we can have… we can overcome some emergency situation. So, we have a backup image for the operating system, and that way we can restore. However in the case of physical damage or radiation, I mean the radiation problem, yeah. So I need to just pray. Actually, we just have one other experience, we applied our software on International Space Station already. And so we have some reference around. However, in the case of International Space Station, the data was similar to our data center. Similar kind of runtime environment and then from ground and then we are, we can push the code and we can get the data we want. So the concept is still similar, but in the security case, it’s more challenging.

So just my viewpoint, and what can be done next generation data center or environment. So I think that I’m not sure what that will include, some edge, end edge based micro data center. Just imagine that, I mean, that the future of the data center can include the edge and include the space part. So yeah, this is just a good opportunity to think such a kind of architecture and the use case, I think.

Blake
I think it’s amazing to see the direction that the computing on Earth is taking between cloud computing and edge computing. We’re really seeing a broad distribution of physical hardware across physical distances and land. And it just makes me think of how people have really stopped using desktops for the most part, because a laptop is all you need. Any higher level or higher intensity computation, it’s all done somewhere else. So I really think that this kind of work is going to change what we think of interfacing with computers is at all. Like, I really think that this is a very interesting direction. Happy to have a chat with somebody so involved. So if I could talk a little bit about the innovation, the mindset that goes into a project like this, because edge computing, as I just said, I think it’s on the frontier of computing in general. What do you think drives that kind of innovation? What do you think drives the idea or the incentive to say, hey, we just did this on the Mayflower, let’s just do it in space now. What does it take to have that motivation?

Minsik
There are a lot of activities in the space area from the commercial companies, and there’s already a lot of technical innovation there. So, we never imagined the kind of network connectivity in space. We never imagined, kind of satellite based or Mars based or Moon based, that kind of infrastructure part. However, now, we already have some kind of network environment in space, and so satellite can use the Wi Fi work 4g or 5g network not only for this space area, in many industries, just to apply the modern IT technology autonomous car or electric car is the main trend in that industry. Mayflower is similar, right? We can just consider the car age or kind of the compute or kind of robot and that can have a kind of brain, some IT enabled brain and they can do more intelligent processing for the drivers. And in space industry, there are so many opportunities to utilize the current IT technologies. So we don’t have to invent technology from scratch, we can just apply that. Once we can have this kind of emerging IT technology in some space use case, we can apply the same thing on other use cases. Yeah, that is our motivation when we start this work.

Blake
And you were saying that you are able to use some of the innovations that other people have produced or the technologies that other people produce to help drive your own. I think that’s something that this project is going to do for a lot of other people. Part of democratizing access to space instrumentation is that people will want to learn more about space, want to get more involved with the space industry, and they don’t have to launch their own rocket or build their own satellite, they can use the stepping stones placed down by people like you to further themselves. And I think that’s fantastic. So I have my own personal interest in edge computing. So I’ve been excited to talk about that. But I think part of this project taking place in space is more of a show of force for edge computing. But I’m interested to know, has this project gotten you personally more interested in the space industry? Or have you always had an interest in space? Is it your interest in space that brought you to edge computing, or your work in edge computing that brought you to space, I guess is the question?

Minsik
Yeah, so actually, I majored in mathematics and information security, and then worked IBM for like 15 years as an architect. When I joined our group, I realized that we can do some interesting things to further edge computing use cases. In my case, I got opportunity to work in this space area. So I’m also learning a lot about this industry.

Blake
I’m very excited to see what the face of computing is going to look like in four or five years as more and more projects like this and people like you get involved, I think that we definitely have a lot to look forward to. I did want to ask, just as far as some inspiration to people who want to follow a similar line of work as you, and I assume that there are many, because any work in computing is obviously highly transferable, you can work on autonomous ships in space, and anyone needs a computer. So do you have any advice for people following a similar career path either in computing or just kind of searching around for a field that they think they can be happy in and be productive in?

Minsik
Yes. I recommend that now kind of machine learning. When you consider autonomous or some edge computing environment, it will be critical to understand such a new trend, especially AI and the environment. I recommend the people have some skill, and then start in the AI industry. And that will enable and bring a lot of good ideas, and then maybe they can find their career paths, I think.

Blake
So keep on top of the trends. Know what’s in.

Minsik
Exactly. So five years ago, I didn’t see any data scientist, as I remember, I mean only a few. Right now, new trend, AI part and AI ML part. So if we have some understanding for that, then use case that will be very helpful now so many new emerging technologies, right? And it’s really hard to catch up. I mean, it will be good to check the latest in the technologies. And then, I’d like to encourage that people to pay attention to the latest trend, and then some technology. Then they will have more idea for their work.

Blake
There’s new opportunities emerging everywhere.

Minsik
Right. There’s no barrier now in any industry.

.

Blake
My conversation with Minsik elucidated something magical about the seemingly sterile field of edge computing design: the way that Minsik is sending computational work into space to lighten the load on the ground computers parallels the way that he’s sending his expertise to new programmers, who soon will no longer have to overcome the financial, technical, and gravitational barriers to experimenting with space. By connecting a satellite only 10cm across to a global network, Minsik’s team at IBM will be connecting young programmers to a space tech community that used to be –quite literally—a world away. That is truly a “stellar” achievement. Thanks again to Minsik Lee for chatting with me and thank you all for listening.

We hope you enjoyed today’s segment. Please feel free to share your thoughts over social media and visit Longitude.site for the episode transcript. Join us next time for more unique insights on Longitude Sound Bytes.