Red Hat OpenShift In the Clouds | Red Hat Livestreaming, 7 Apr 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: In the Clouds (E22) | Edge Computing in Space

Description

In this episode of In the Clouds, we shift the conversation into orbit as IBM Distinguished Engineer and CTO Space Tech Naeem Altaf joins host Stu Miniman to tell us how together with NASA, cloud and edge computing services and technologies are being used to solve problems on the International Space Station. Join us to hear how the cloud computing technology and edge computing project collaboration began between NASA, IBM, and Red Hat and discover some of the significant benefits of edge computing for space-based systems and why they'll be important for the continuing exploration and exploitation of space.

Want to learn more?
https://www.redhat.com/en/resources/edge-computing-in-space-brief?sc_cid=7013a000002qG16AAE

Try Red Hat OpenShift: https://red.ht/3Haxapr

Follow us:
Stu Miniman https://twitter.com/stu
Naeem Altaf https://twitter.com/lahorite
OpenShift https://twitter.com/openshift

Subscribe to Red Hat's YouTube channel: https://www.youtube.com/redhat/?sub_confirmation=1

About OpenShift:
Red Hat OpenShift is an open source container application platform based on the Kubernetes container orchestrator for enterprise application development and deployment.
OpenShift on redhat.com https://red.ht/3JBaOia
OpenShift on GitHub https://github.com/openshift/

#RedHat #OpenShift #HybridCloud

A

B

My let's get into the big topic of open source, something that we.

A

C

B

What, let's say.

C

Has kubernetes ecosystem really boom.

B

Welcome to another episode of in the cloud, I am your host stu miniman and we could say that this episode is going to be out of this world. Yes, we are going to be talking about a special edge computing project uh on the international space station, something that ibm and red hat work together on. uh Obviously uh many of us in tech are are space enthusiasts, uh so uh I, for one, am really excited to be able to talk about it.

B

It's a really a great example of what ikt can do when, when teams have reliable infrastructure uh in the software to go along with it, uh there's actually there's a reason why red hat has a new brand campaign. So uh if you watch television, especially if you watch sports, you might see uh this new advertisement that we are going to share with you here, uh quick 30-second ad, and then we will be back uh with the program. So uh please enjoy this advertise.

A

Test commencing.

C

Good code should work anywhere, even in outer space.

A

Water safety test commencing in five.

C

When we know our apps will work in any environment, my team can focus on our code. We did an extra round of debugging right yep. The code is clean. Okay, here we go. Everybody.

A

Houston water analysis was a success.

C

This is what connecting your clouds feels like.

B

Awesome, so uh you know, as I mentioned uh before, that uh you know space activity uh when I look at cloud computing, what we've been doing at red hat with openshift, uh we talk about having our code run anywhere edge. Computing, of course, has been uh a big effort to be able to make the footprint fit and talk about. uh You know some of the some of the challenges you know edge computing we're going to the far edge when we go to something like the international station. So I'm really happy to welcome to the program.

B

Our guest for today is naim altif. He is an ibm distinguished engineer and he's also the chief technology officer for space tech naeem. Thank you so much for joining us and uh gotta say I. I love this. The space setup that you've got behind uh next to you there with you, know the rockets and satellites and the likes uh thanks so much for joining us. Thank you so much for having me all right, so so name uh you're a distinguished engineer at ibm. So you know distinguished engineers and fellows. You know.

B

These are people that have you know broad experiences and done a lot. um Tell us a little bit about your journey to space tech. I mean you, you weren't, you know somebody started out you're for more of the software side, so help us understand a little bit your journey as to how you got to today sure.

D

So, like you mentioned like uh we all in the you know, technology sector, weird space enthusiast, many of us are so I think few years ago I got the opportunity to work at the seti institute as a volunteer. We were working on some interesting projects and from there I got this. You know visibility where, as you know, most of the industry today is going software defined. So, like you said, I don't have a you know, aerospace background, I'm a computer scientist, uh so it's like okay, you know, programming is everywhere.

D

Code is everywhere, so it means that we can also be in this uh in this sector. So that's how I got interested and started learning about what are the different segments in this space where vs software developers can play in and that's how you know all these ideas came into picture and we'll discuss more about these things.

B

Yeah so uh you're at ibm, of course, ibm deals with companies not only around the globe, but now now you're dealing with solutions in space so bring us help set the table for us. You know working with nasa. What was nasa looking to do? How did ibm get involved? What were some of the you know the challenges that that that they pulled you and the team uh into sure.

D

So uh when we started looking at, uh you know what interesting things we can do in in space, so right now the buzz in the industry, as you know, from we went from cloud computing now in in the era of edge computing, and the idea was: how can we extend this edge computing into the orbit in space and uh the target was international space station international space station, as you know, is a laboratory for many. Many experiments happening there and we we can judge. We know we have.

D

We know the communications we have on terrestrial networks. If you are at some forward location, you have latency, you know bandwidth challenges if you're in space, those magnify those challenges, so we reached out to space station national labs and nasa, and we like, okay, what kind of projects you're running there, which are heavy on compute heavy on data, so you can do some match computing for you and the project which they mentioned was one of the projects was dna sequencing and they mentioned that.

D

You know it creates almost half a terabyte of data on each one and now that data has to be, you know, brought down to the uh to the ground through through the ground stations and the way you know just a bit of background uh when you are in the low earth orbit which, where the space station is around, I think 250 miles up or kilometers up. You are moving at 17 000 miles per hour, so your touch point with a ground station.

D

Any given ground station is only seven minutes and you're talking about a radio frequency. So you know there are several bands but you're not talking about gigabits of transmission, so in space station in the any satellite in the lower earth orbit. If it wants a much more frequent, uplink downlinks, it actually goes up to the geostationary flap satellites and nasa has a constellation of them for the data relay.

D

So the data goes from the space station up in the geo stationary orbit and from there it comes down to the ground station and the reason for that is, if you have dish at home like dish, network or drink whatever. You know you have you probably have noticed once you put the dish at your home, you don't have to go out and keep on moving the dish, because earth is rotating satellites in geostationary orbit. The satellites are in mo in exact motion of how the earth is revolving.

D

So that's why you have these dishes, which can point to your telecommunication satellites. So that's how it works from the space station goes up and comes down. So while we were uh working with them, so we started how about if we uh perform this computation on the space station. So we work with the uh you know with the nasa try to understand. You know what needs to be done and it's a it's a bit complex.

D

uh uh You know uh problem because uh when you're talking about dna sequencing, there are multiple steps involved, you have base scaling, you have the demultiplexing, you have the alignment and then you have a general report. So we took that code. We containerized that code and we wanted to get the openshift platform. We want to put the containers in space. So at that time, two years ago we had the code ready container from the uh from red hat openshift, which could fit on the single node. So we took that code.

D

We containerized all the code and we didn't put this. uh It actually went up last year in april. uh So and since then we have like three or four runs. So basically it's running in production. Now and and again the goal was minimize the downlink. You don't want to downlink half a terabyte. Now, for example, when the report is generated, it's literally one meg file.

D

So so again, so you did all this computation there and because the principal investigators who are requesting this run, they just want a resulting file. They don't care about all this raw data and right so and and and and the genesis of this project was uh they're looking for microbes inside the space station, so basically the astronauts they take a sample from the surface of the space station and they run through this device, which generates all the data, and we pick it up from there and generate the final report.

D

So these has huge implications for the future. For example, you know we are hearing a lot about. You know, probably watching the rtms program going back to the moon. The sls is already on the pad to go up right uh for the uh and then you're going to the mars, and- and you know for all these deep space missions. Imagine you have these robots and you have these uh dna sequencer these devices. So these robots just like how mars does a mars robot does right rover.

D

It goes and it pokes into the ground to pick up the sample. So what if it would pick the sample- and it has a dna sequencer there and it can run through the sequencing and it maybe it can find life somewhere. So that's the whole idea about edge computing and edge computing in space.

B

Yeah, no, it's it's! It's fascinating! uh When you look at this, uh it's one of the patterns we look at for artificial intelligence. Often you might do your training at a centralized location. Like often the cloud is where you need to kind of build the models, but you need to be able to deploy where I'm at the data source is so the example we've used in the industry for a long time. Is you know if you look at cars?

B

If I have a car and it has a lidar, I can't wait to go back to a public cloud because I will have hit you know.

B

The thing before I get the data back space you know absolutely is is a phenomenal example of that, um as you say, just on the iss there's limited bandwidth, but right if I'm going to the moon, if I'm going to mars, if I need to be able to analyze data- or you know maybe look at certain safety concerns, where I would be testing things that would be so important to be able to do that locally um question for you. You talk about the maturity of kubernetes.

B

You worked with our team at red hat. um We spent a lot of time over the last couple of years shrinking the footprint and the amount of resources there. So maybe you could talk to a little bit the intersection of the requirements that they had and uh the the maturation of the technology to reach. uh You know the successful solution.

B

D

No, so that was a and it's a great uh partnership with the redhead and uh and shout out to the team. You know at the crc and we'll talk about the micro shift as well, so there, so the goal was to reduce the enterprise uh open shift version to a single node two years ago, which I know it exists today. So so that was the initial goal then, from there the next step was okay. How can we shrink even further?

D

Because when we are talking about edge computing devices, we are talking about devices like uh raspberry pi stuff, like that, can we have a version of an open shift which can fit on this device? And we can you know and also beyond you know, besides these uh versions of open shift, we also have, as you know, apartment to run the containers. So we are using a combination of we how it fits where we use portman. What is the lighter weight of the uh you know?

D

The uh open shift, so that's how we know we work together because the uh the target needs for we need a lightweight platform, because the footprint of hardware is not as big as you can. You know expect on the industrial networks, so that drove all these uh requirements with the team, and uh it was a great experience working with other team.

B

All right, one of the other concerns also is you know: do I have the skill set to be able to take care of things so how many red hat and ibm people did we send to the iss to be able to set up? You know, configure and manage this whole environment.

B

This one did we actually have uh have an employee from uh uh up on the iss or you know what was it somebody that was trained.

D

No, no so we have the uh actually. We have access to the systems from the ground, so we were able to like ssh into the systems and basically the flight systems were built here in alabama, so we uploaded everything over there wrapped up system and once it went up and it took like couple of months to go live at that point. We were able to go and you know start the uh the whole sequence so, uh but at this point I mean minimal right.

D

I mean the initially with ibm engineer, redhead engineer, I'm making sure the system is all prepped before it launches and after that, it's just a matter of running the script.

B

Yeah um yeah that's phenomenal and that that's the you want it to be able to manage things. You know out of that, but you you don't want to have to have the resource touch there. Automation needs to be involved. You know remote management, all that sort of sort of stuff. um What else anything else we should know about. You know the kubernetes solution uh on the space station.

D

No so, like you hit the right point fully automated, that's the key thing, because you you have like at any given time between three to four astronauts on the on the space station and there every minute is scheduled. I mean I have seen those uh schedules. I was in the iss uh headquarters in dc and I was just surprised to see up to a minute that they will do this. They will. They will sleep for this minute they will eat now. It's like everything is scheduled for weeks.

D

So so we had to make sure that the platform we put in can fully because, like I mentioned, there are multiple steps involved, it's a complicated process, so we need a platform which can take care of which is resilient, and you know in case of any failure. It can go back and start, so we shouldn't have to ask somebody whether on this space station or on the ground station, at the nasa to go and trigger this again, a fully automated process.

B

That's awesome um you mentioned uh microshift so uh for those that aren't familiar micro shift is currently a open source project uh to be able to take. uh You know when we had open shift, uh we can do a single node um of of openshift, so kubernetes on a single server. But there's still, you know a minimum requirement for how much resources are available. Microshift actually isn't full kubernetes there's some things we need to strip out.

B

We need to be able to make it to fit on a smaller footprint, so where's where's, microshift intersecting, with uh with what you're looking at on the space programs.

D

So so that's a good good question, so our goal is to introduce a containerized platform for the space industry and, as you mentioned right there, there are a lot of uh restraints here because of the the footprint of the compute.

D

So actually uh there's a there's a very, uh I think, the only conference for the small set category, which happens every year in utah and we just submitted, which got accepted so this summer we will be actually presented, presenting microshift to the developer community of the satellite and it's a huge community, and we want to introduce them that, as you are building schools, universities and the small set programs. Take a look at this platform which we will talk about what we are doing in future with this platform on our mission.

D

But uh but that's the idea right, I mean uh make it available to the developers and let them play with it and see how they can. You know orchestrate from their different kinds of workloads, because the thing thing about this are the the data centers we have on the ground today.

D

These data centers are planning to move in the orbit yeah. It sounds like little bit, but, but that is true: they want to move these data centers in the orbit, so you will have basically a cloud will be running in the orbit. So all the things which we care about the multi-tenancy right, how can the platforms handle multiple workloads, the uh sas kind of solutions they all will be going into the space we just have to be?

D

We can't just go over there physically and do some things you have to keep that in mind. While we are developing these systems and solutions.

B

Yeah I I had too many years. uh You know watching the data center space. The thing that comes to mind is like well, cooling should be less of an issue in space because we should be able to you know, vent out heat rather than uh we're worrying about uh other things. There um yeah, but maybe we speak a little bit about the endurance mission and the cube satellites and where people can learn more about that sure. So.

D

uh When we started uh looking into the uh you know into the space area, one was one thing was to: uh how can we also launch our satellite, and this is purely for research and educational purposes, and uh so so we started working on that and we wanted to use the latest and greatest technology available among us, ibm and red hat, and at that time micro shift was still in very it's a. It was a thought process right.

D

So we've worked with the team over the year and finally, we have open source version of micro shift. Actually this will be launching on june 1st net 30, which means you know because of the weather. It can get delayed few days, but the target is the june 1st, the largest cubesat, which will be, uh which will be running on raspberry pi, and the idea over here is, as I was going through the process of. How can we put a satellite? If you do a look for a checklist?

D

It is huge and for right reasons, because you know it you have to go through so many steps and approvals, and I mean if you you guys, can just google nasa 101 cubesat, and you will see this like 200 page book guidelines, which tells you you have to follow all these steps. So I was like okay. If, if we're living in these, uh uh you know modern part of the world, it is so hard. Imagine the developing nations or underdeveloped countries across the globe.

D

It will be almost next to impossible for them to launch something or have access. So the goal was okay. We will launch this thing and we will make it available to every school kid on the planet. It will be available so uh so right now we are starting with ibm ptek program, which is the schools they work with all over the globe, and I think the uh the link is shared. So so schools can register and we will be.

D

We will give them the apis, everything they can write their code in python and they will be able to submit the code and the code will actually go on the uh on the cubesat, execute there and and come back. There are variety of sensors on this cubesat from you know, from all the positioning there's like, I think, 20 different metrics for positioning vector you have the uh all the metrics played for the battery for operational point of view. You have the metrics for the solar panels.

D

You have the the heat information, then you have the compute over there. Where we can use commands like you know, like nmon, we will have podman there. We will have microshift there. We have ai binaries there, so we can detect. You know, because you have a camera on the cubesat, so it can go and detect stuff.

D

So we have variety of sensors variety of metrics, which your code can play with and do some edge computing computation and then once you one last thing, once you uh execute the code and the results come back, we will issue you a certificate which will be uh created by basically the stamped by quantum computer. You know the number you get will be quantum community generated.

B

Yeah naima, you know I I know I was excited when uh you know I could have my name. uh You know on something that was like going up in a rocket and like nasa, send you a little certificate, but what you described there, you know so many home enthusiasts. You grab a raspberry pi, it has centers. I can do that, but you're removing the hardware and it's in space. So how cool is that the kids get to play with this? They get there. I. I would hope that this will.

B

You know just increase people's interest in stem. uh You know we we need, you know globally to be looking at. uh You know the the the the opportunities of space, so uh that is that is so awesome uh to see the activity. um You know something we expect. You know a company like ibm to be involved with. So it's great to get back on that. um You know you talk about how everything gets all planned out. um You know I I watched lately, it's like up!

B

Well, you know we have the rough date as to when the iss will. You know, be decommissioned and come down. So um what can you tell us about? You know future activities, you know: are we planning for future space stations? Are there other? You know, projects that that you can share that. uh You know ibm and red hat are involved in.

D

Like cool yes, so I mean space station has been amazing right and it's coming end of life at 20, 30., and actually you know there is a there- is a graveyard selected for where it will go in the ocean. It's the west of australia right that was really cool. I learned recently about that. So so now, as space station is, you know uh coming coming down our end of life.

D

There are news companies who are emerging, because this is what has happened in last couple of years and uh and a huge you know uh stuff, which uh the industry which has been changed, is by, of course, by spacex what they have done. So basically, the the commercial and the private sector is coming together with the government sector right they're opening up- and you will see there is a there- is a flight going up tomorrow, right uh ax1 for the private astronauts to go on space station.

D

Then there is the crew 4 going end of this month, so I mean again going to the space station. But now, since the uh the private sector has, you know, come forward, there are many companies like. I think there are three or four who got the awards to put these space stations the next nursing space station.

D

So if you think about the space station, if you think about insight it, you know it has lots of sensors, it has lots of research equipment inside where they're doing research for microgravity. For you know, looking growing plants, the food stuff, I mean watching the health of the astronauts and doing you know a lot of imagery from on the planet, all that kind of stuff, but inside to do all those things you need. You need id infrastructure, you need compute, storage and network, and that's what these companies are.

D

You know we are in discussions with few and talking about okay. What will be the next generation platform, which is the whole stack right? Your uh your infrastructure, your platform, your tasks, your sas, the whole stack. How will that fit in that uh space station? And then you know the rules will be a bit different. How you interact with. Basically, the idea is the space stations will become because they want to commercialize this stuff, so they will become like labs right. So you have different kinds of experiments.

D

You want to run so you're on the ground and you know a cloud of your choice and you know from their cloud. You say you know I want to access something. I want to run some experiment on that, given space station, so that also brings in the picture of this ai sec devops. You can add so many acronyms right so the whole pipeline. Basically, so that is a huge uh discussion. Okay, how can we build these pipelines where users can build their code? And that's where you know the uh the cool thing about openshift?

D

Is you you write once you can push your code anywhere and that's our messaging is that we have a. We have a platform whether you're in your data center or you are in some somewhere in the cloud or you are at the edge somewhere on the test network or your edge happens to be in space. Your underlying platform is the same, so me as a developer. While I'm writing my code. When I push my code, it doesn't matter whether it's called in space edge on the ground, because the platform underlying is the same.

D

So so that's one area. The other area is: how do we build these pipelines to push and then the transmission there are challenges the network? You know the disruption between in the orbit. How do we make sure that it's resilient, that whatever I pushed to this far edge? It's the same thing if it did not got compromised, especially when we are talking about the uh machine learning models? It's very, very sensitive.

D

You know, because you're relying on the output of those models, so you want to make sure the model which you developed is the same model which is running over there so yeah. These are all the different things I mean you will think about the platforms for software development, the your data pipelines, how you're going to push security is huge area how you're going to secure the communications so- and this is basically data center- is now going from the terrestrial into the space.

B

Yeah, and even I I think about you- know public cloud architecture name when you talked about you know the geosynchronous orbit, you know, do you expect to see a public cloud announcing a region? uh You know on the geosynchronous satellites because it's probably still going to be a little bit while uh before they have an availability zone on on the moon. So you know with the geosynchronous orbit, be the first place that you know I might have certain cloud services available there.

D

Most probably in the low earth orbit and the reason is the companies are talking about putting a constellation. So you have a ring of let's say you put six or eight of these mini data centers and they will be connected through optical connections in a ring and then based on that you can write information among each other and push down whoever has the best site to get to the ground station. So I think definitely we will see these uh edge locations in space in few years.

B

Wow, um it was interesting. I attended uh back in 2019 when we still had physical events. Amazon held a conference that was uh you know. Space was a big piece of it and I I geeked out. I went to a whole bunch of sessions and I talked to some friends and I was like oh, I wish I was in the space industry and they're like stu compared to the I.t industry. The space industry actually moves kind of slow. So I'm wondering if you know you've got a great viewpoint of this.

B

You talk about all the planning. You talk about the 200 pages. You know today's day. You know I I want to be able to respond fast, spin things up do instant. So where do you see kind of space and I.t intersecting when it comes to the path of innovation?

B

D

Think on the on the on the uh development of physical infrastructure, it's going to be slow because you only have one shot and that's why you know I mean in space industry one launch failure, you're almost you're out of luck. Your company is done so that part is hard to move faster, but the things which will move is faster. Is the software defined and that's what we are trying to focus? Okay, you know. Yes, you you guys are expert in building these spacecraft satellites.

D

We are expert in the you know, on the software side, how can we come together to create these software defined satellites again? Think about it? The the challenge of solving data on the on the ground. I think this is far more advanced than the type of data you will have in space. I mean think about you know all the social media, the amount of traffic flows through networks every day you don't have that much traffic, so I think we have pretty good idea of how we can build these kind of platforms which are resilient. Now.

D

The challenge is how we can fit those in a smaller, tighter space. That's the challenge, and this and the industry is moving towards courts course means you know compute off the shelf so before it used to be proprietary, fpgas stuff like that.

D

But now it's opening up because there are companies who are building these uh enclosures, which can protect from radiation radiation is the biggest thing to protect your computer from so so they will be like radiation hardened enclosures, where a compute can your regular compute can sit in and it can do stuff right so uh so these are some challenges, but I think that the aspect of software defined for the space industry that will speed up things.

B

Yeah, um I I'm curious, you, you live down down in the austin area and we've seen you know. Texas just seems to be the hotbed, uh for you know space. uh You know these days. um Yet in the last two years a lot of the stuff has gone remote. So you know speak a little bit. If you could too, you know what's happening in texas versus you know the global participation uh in all of the space activity, like your team, is it you know, is it a globally uh dispersed group or you know or is?

B

Is there a core team? You know in one of the ibm locations.

D

So these days you know everybody is in remote locations because of you know the pandemic or last two three years, so the core is team is in austin texas, but we have participants from almost all over the world because you know we, we have a thing called in ibm uh the academy of technology where you go and create initiatives and people worldwide join, try to help you to build those different kinds of solutions.

D

So it's a family and we also had participants from the red hat of course, so we all working whoever can contribute from wherever they are in their capacity. So that's where uh you know how we are working now, if you talk about you know, what's happening in this region right, so you know we have uh boca. Chica right has a very cool name in in south texas. Over here we have starship right and probably guys saw in last year when they did their test with starship. So I think that is.

D

uh We are very looking forward to that when actually the starship launch happens and a huge and a lot of things are happening. You know they're talking about building the uh the next generation of spaceport here in the uh in the south, but that will also bring modernization of your command and control data centers and that's where again, our platforms, our technology will come in so there's a lot of room as software developers uh with little or no space background all week. All we want is your uh enthusiasm and your.

D

You know your curiosity about space. I think there's plenty of room.

B

Well, I guess that that that's a great setup for my my final question is you know where what are the skill sets that that people are looking for um for those people that work in tech that are enthused? You know what are people hiring for what what are the areas where uh you know that there's real need to be able to uh you know I might not be the person getting on the rocket, um but I can help uh you know, build the technologies around uh what's happening there.

D

Sure again, let's I mean just just to look at the uh the recent examples right. So look at how the the booster comes down. That's pure programming, how it's coming down! You know that, that's just you know amazing to watch it's. Sometimes it looks like you're watching a movie, and then you saw when the uh the uh dragon which took the first uh crew to the space station. You saw the consoles right, the linux space or windows based consoles inside so yeah.

D

As long as you know, hardcore programmers, there is plenty of opportunities and then, if you have you know again, if you want to be in this aerospace or space industry, you need to understand a bit more about how the dynamics works here in the space. You need to have some ideas about, and but if you are a good programmer, I don't see you know how you can fit here, because they need lots and lots of programmers.

B

Awesome well name, thank you so much for sharing. uh We we definitely. uh We we've got a chat group here at red hat, that you know, watches all the space flights and uh really is interested, and you know a great collaboration between ibm and red hat to make these solutions work. So thank you so much for joining us. Thank you so much my pleasure.

B

All right so uh hope you enjoyed this special uh session uh about space. You know definitely a topic. uh I I'd love to hear, uh hear plenty about uh just programming note. uh We will not have the program two weeks from now. We will be back uh in four weeks. We've got a lot coming out. uh If, if you look at red hat red hat summit uh is happening there in may so look for lots of announcements. um You can join the event. It is a hybrid event and registration is open.

B

So uh thank you so much for joining us uh for your journey in the clouds. I'm stu miniman, we'll see you back in four weeks.

B