In an exclusive interview, Maj. Gen (Retd.) Clint Crosier, Director of Aerospace & Satellite, Amazon Web Services (AWS) shares insights on the ongoing shift towards cloud, the era of AI & ML, on-board processing, in-orbit computing, sustainability initiatives on Earth as well space, and more.
“What mechanized manufacturing proved to be for the industrial revolution, Cloud will be to the digital revolution. Companies, organizations, and governments that don’t take advantage of the advanced technology available with the Cloud won’t stay competitive or relevant over time”.
New space ecosystem today is evolving at an exponential pace, especially with trends such as sensor miniaturization, faster analytics, rapid connectivity, IoT, and extensive downstream capabilities. Where do you think is the sector headed and what would be the role of geospatial in it?
The reason we have rapid connectivity and faster analytics is because of high-performance computing, Artificial Intelligence, Machine Learning, and global cloud infrastructure. All of these are very much connected. Coming to future trends, there are a few things. First and foremost, we see a continued movement of space and geospatial organizations and companies to the cloud.
Three years ago, I would hear from companies that they didn’t really know what value the cloud provides. And then, within a year they recognized the value that it offers. Now the sentiment has shifted towards swiftly moving to the cloud. This is going to continue.
Interconnectivity is the other trend. The history of space in the realm of geospatial has been that satellites only communicate with one ground station. Satellites in Geo orbit didn’t communicate with satellites in LEO (lower Earth orbit). This led to a stovepipe set of capabilities.
Now, increasingly users, such as those who travel around the world, don’t really care which network they are on, or which system. They just want everything up & running and functioning. The same is the case with space.
AWS is working with companies to create the connectivity layer to enable IoT on Earth. AI has proven to be extremely valuable, and Generative AI built on top of it will lead to an exponential leap.
There are immense possibilities for using Gen AI, whether it is new concepts for spacecraft and rocket design, or developing faster ways to train large language models on geospatial data to identify things that the human eye can’t pick up. Going forward, these would be some of the key industry trends, and the cloud plays a big role in all of these.
AWS has been a pioneer in this new paradigm of Ground station as a service. How do you see it beneficial to startups, especially in the emerging markets?
What I love about AWS Ground Station is that it is basically built on the fundamental premise of the cloud. Instead of spending your limited funding and your limited capital on building out infrastructure, use that instead on your actual mission and let AWS provide you with global infrastructure. That’s the basic premise of the cloud.
When you’ve got a small company that’s preciously guarding limited resources, they don’t want to have to build the infrastructure of a ground station network instead.
So customers can take advantage of the AWS Ground Station pay-as-you-go model, pay only for what they use, and then instead of that large capital outlay, you can put that back in the exquisite nature of your sensor or your data distribution capability. And so that’s the value of Ground Station as a service.
What is the AWS vision for a more sustainable world ahead?
AWS and Amazon are absolutely committed to sustainability. Through the Amazon Climate Pledge, we support climate and sustainability. When we look at AWS and we look at space and geospatial, I think there are two interesting things. The use of space to support sustainability on Earth is the first.
There are companies such as Orbital Sidekick that use geospatial data and AWS to monitor oil and gas pipelines around the world to detect and monitor any emissions from leaks or damage, especially methane gas exposure. This is an important focus for the energy industry.
In India, there’s a company called SatSure that uses Earth Observation data and Amazon SageMaker to help farmers identify changing conditions. In the UK, we work with a company called SatelliteVu that uses geospatial data to identify thermal emissions from manmade structures.
Digital Earth Africa uses artificial intelligence and machine learning on geospatial data to improve food security and agriculture production in underrepresented regions of the world. They tell us that with the help of AWS Cloud, their analysis predictions get up to 800% faster. Space geospatial and cloud capabilities are together enhancing the sustainability of Earth.
There’s a need to ensure that we keep the space sustainable by not polluting the orbit with unnecessary debris. For this, we work with companies such as LeoLabs that have built their entire space traffic management and space collision avoidance capability on AWS to provide real-time monitoring of objects in orbit.
We are also working with companies such as Astroscale in Japan who use AWS to identify ways to keep satellites in orbit longer, and being able to maneuver them when they run out of life to avoid the possibilities of debris. Active debris mitigation capabilities are crucial for space sustainability.
Earth Observation and Satcom are the two biggest markets in the New Space economy. With a plethora of space data on the Cloud, would there be a constant need for the future of secure and encrypted communications everywhere?
The biggest telecom organizations around the world virtualize their networks on AWS to boost innovation, reduce costs, and improve resiliency and redundancy. This is already happening with terrestrial communications. The same trend will continue as we move towards satellite communications.
Last year, Thaicom, one of the largest Satcom providers in Asia, announced they are working with AWS to develop its cloud-based TV broadcast distribution platform. They are virtualizing their network software-defined capabilities so they can increase innovation, decrease cost, and increase the resiliency of their customer base and network architecture.
Earlier this year, we signed an agreement with Eutelsat OneWeb. Eutelsat OneWeb is going to move its mission operation centers to a virtual cloud-based network, and we’re collaborating to improve the delivery of satellite communications capability and cloud-edge computing edge services.
In terms of secure communications and capability, one of the things that I am really proud of with AWS is security – it’s number one in everything that we do. There’s no other cloud organization in the world with the size and scope of what we can commit in terms of global resources and expertise.
We are going to continue to see satellite communications move to the cloud for virtualization to take advantage of the built-in security on the AWS network.
It is often said that data is the new oil. While that may or may not be the case, but it certainly is an invaluable asset for space-based applications, systems, and solutions. With petabytes of data on the Cloud, what would be the role of AI and ML in the future of analytics?
We have seen that AI and Cloud go hand-in-hand and grow together. The two are inextricably linked. AI and ML get more powerful due to Cloud and vice-versa. So I think you’re going to see those continue to grow together.
When we look at space and geospatial, a lot of studies point out that there would be five to ten times more satellites in orbit over the next decade. This would also lead to an almost infinite increase in data and bandwidth that’s coming down.
The only way to make sense of that data is via AI, ML, and advanced data analytics. The idea is to take that data, turn it into information, and convert that information into knowledge, allowing us to make real-time decisions.
Timing would be equally critical. We can’t afford to wait for say a week a day, or even an hour to process that data. Cloud, AI, ML, and advanced allow us to turn raw data into quality insights that decision-makers and civilian and military leaders around the globe can use to make decisions that affect people’s lives.
We see in-orbit processing, edge computing, and the scramble towards quantum communications. What would be their impact in the coming years, and how do they align with the AWS future roadmap?
They align very well. Over the last decade, we have been working on extending the cloud coverage. The basic vision of AWS is that we would bring the cloud to your data or applications, rather than the other way around.
And that’s essentially the birth of edge computing. We want to help customers reduce latency, increase efficiency, and provide the cloud when and where they need it most. That’s the key.
Our customers have missions spanning space, aerospace, and geospatial. For instance, when we talk about missions such as the NASA Perseverance Rover, it spans a 300 million-mile journey and processes hundreds of images from Mars each day.
By using AWS, NASA Jet Propulsion Laboratory (JPL) can process this data faster than ever before. The increased processing speed helps NASA JPL make faster decisions on the health and safety of the rover and prepare for the next day’s activities.
For the growing missions in the LEO economy, we will be building satellites and launching them in space.
We will also be into satellite servicing and space robotics. If we look at the Artemis Accords that aim to put a man back on the moon, and also the first woman, along with creating a lunar infrastructure and cislunar economy.
For all of this, we cannot afford to push all that data to the Earth, process it in the terrestrial cloud, and then push it back into space. There is a need to move the cloud where our customers need it and they would be requiring it in space.
We did two interesting things over the past two years. One of them is on-orbit testing using an AWS Snowcone, an edge computing device that does ML and computation. It’s designed to be rugged and lightweight.
The device underwent a seven-month NASA space flight certification, process shock testing, vibration testing, and the whole nine yards. After that, it got certified for launch on a rocket to the International Space Station (ISS).
Once it was installed on the ISS, AWS and Axiom Space started conducting experiments processing images from onboard scientific experiments.
Together, we tested how the device allows us to process data on the space station, instead of pushing all that data through limited bandwidth, and limited pipes back to the earth. As a result, we saw a significant increase in the ability to do on-orbit experimentation, science, research and development more quickly and efficiently.
Next we built our own purpose-built hardware software package, and we partnered with industry partners D-Orbit, and Unibap to conduct a software experiment on an orbiting satellite.
When pushing satellite data to the Earth and processing it, inevitably some of the images are unusable because of cloud cover, or the images aren’t the required ones. By pre-processing images onboard the satellite instead, we discovered it’s possible to eliminate downlinking that unusable data.
We found out that 100% of the mission could be met with 42% less bandwidth for the downlink. That again increases the efficiency of what satellites can do, allowing customers to fetch more data, which again needs more cloud-based AI and ML. This space data processing cycle is an important one that relies on the cloud.
Watch the interview here:
