Max Peterson, vice president of the Worldwide Public Sector at Amazon Web Services (AWS), joins Yahoo Finance Live to discuss how AWS is using the cloud to collect data in space and increase the speed of transmitting data from satellites back to earth.
AKIKO FUJITA: Well, Amazon is taking its cloud services to new heights. This as the cloud computing giant announced today it is successfully running compute and machine learning services in space. Joining me now to discuss the program is AWS Worldwide Public Sector VP Max Peterson. We've also got our very own tech editor, Dan Howley, joining in on the conversation.
Max, this is the first time AWS has operated a suite of these services directly onboarding an orbiting satellite. Talk to me about how this is in line with that larger goal of being able to process the data in space in a more efficient manner.
MAX PETERSON: Yes, good morning. Thanks for having me. Well, I tell you what. It is really exciting because it is the first time ever that AWS has now been running machine learning model modules and analytics on board an orbiting satellite. In fact, we've been doing it for 10 months with our partners, D-Orbit and Unibap.
And the reason that it is important is because you collect massive amounts of data, and what you need to do is turn it into insights. So an average satellite will collect 100 terabytes of data in a day, about 20,000 high def movies. And the challenge is, how do you get all that information analyzed and down to the Earth in a way that lets you take action?
So with AWS's on-orbit machine learning models, it reduces the image sizes by about 42%, allowing the information to be communicated to the ground faster. And it can actually do image recognition right on board the satellite, being able to detect cloud cover or wildfire smoke.
DAN HOWLEY: Max, what is this kind of purpose for this kind of capability? What businesses, what organizations would traditionally want something like this?
MAX PETERSON: That's a great question. It's really exciting right now because we're seeing an explosion in commercial space operations. Over 10,000 different firms right now are trying to innovate to deliver new capability and new insights from space. So this ability to do the data processing onboard the satellite is essential. And examples would be Earth observation for the purpose of disaster response and disaster preparedness.
Imagine that you can train the model to detect fire or smoke or weather patterns. Now you can do the analysis right in space, using the machine learning capability, and then take action faster. It'll be useful for deforestation and sustainability. It'll be useful for migration and disrupted populations. I think the number of things that we can do are just infinitesimal.
RACHELLE AKUFFO: And so when you think about perhaps some of these public services, governments, et cetera, who want to use these services, any concern about how they might be using this data?
MAX PETERSON: Well, right now, the opportunities to be-- use the satellite information, satellite data, as many people have seen, is incredible for transparency. And so you're to be able to take images and understand where ships are moving across the ocean. You're able to take a look at where land formations are or disaster events are occurring. So I think the opportunity is really to be able to use space in so many ways to improve life here on Earth.
DAN HOWLEY: Max, can you tell us a little bit about what some of the complications were of getting this to work? Obviously, you know, you're running a computer in space, and it's taking kind of orders from there. What does it mean for the company to be able to do that in such a harsh environment?
MAX PETERSON: Yeah, at AWS, we love pushing the boundaries, experimenting and innovating. And in this particular case, it took a lot of work with our space-qualified hardware provider, Unibap, to be able to install the AWS machine learning models and the IoT green grass components on the space processor. Then, of course, we had to work with D-Orbit. They launched this into space. This was in January of this year.
We immediately contacted the AWS component on the satellite, verified that it was live, and then over the course of the next 10 months, started doing the experimentation with the image analysis, with the image compression, and with the various analytics. And the cool thing is, we can continue to train new models with massive amounts of data on Earth in the AWS cloud, and then you just upload the model to do the inference in space. It's very efficient. It drives tremendous cost savings, and it drives incredible speed.
AKIKO FUJITA: Max, you've already got a pretty robust footprint on Earth, right, when it comes to AWS. I mean, to what extent does this represent the growth prospects that you see? I mean, how significant, when you talk about processing all this data out of space, where does that fit into the larger AWS future?
MAX PETERSON: Yeah, you're absolutely right. The footprint on the ground now encompasses more than-- or encompasses 30 AWS cloud computing regions, a tremendous amount of capability on the ground. Many of those regions have ground stations, satellite antennas that now can be used on a pay per use basis by any provider to be able to get that information from space down to Earth, and be able to get models and other information back up to the satellite.
And so it has become more important now than ever that the cloud reaches out to the very edge of work that people are doing. So the very edge of processing data in space or the very edge of high speed telecommunications networks. So AWS is continuing to push the boundaries of cloud computing as far out and as close to the work as possible.
AKIKO FUJITA: It's certainly a huge milestone for AWS. It's good to have you on to discuss this. AWS Worldwide Public Sector VP Max Peterson, and our thanks to Yahoo Finance tech editor, Dan Howley, as well.