UP42: Delivering the power of geospatial insights to solve global challenges
About UP42
UP42 is a developer platform and marketplace for Earth observation data and analytics that help to solve problems ranging from glacier erosion to oil leak detection. The platform offers a library of geospatial data sources and AI algorithms enabling clients to build their own solutions using UP42's API and unlimited processing power. UP42 is a subsidiary of Airbus SE.
Tell us your challenge. We're here to help.
Contact usUP42 built an empowering “geospatial platform,” won planet-scale data crunching power, developed an innovative data analytics dashboard, and gave partners granular revenue sharing on Google Cloud.
Google Cloud results
- Impresses clients with six terabyte computing boost in under one hour with Google Kubernetes Engine
- Meets partner needs with granular revenue splitting solution built on Looker
- Enables planet-scale geospatial data crunching with global data center network
- Inspires new AI and business expansion plans with TensorFlow and Anthos
Making it easier than ever to access geospatial data
The Earth has never been more photogenic. From satellite images taken from outer space to exotic landscapes captured by drones and planes, there's a breathtaking variety of aspects of our planet that are being recorded every day.
These images have become a priceless repository of geospatial data (intelligence about the earth's surface) that drive solutions in a vast array of fields, including wildlife preservation, land management, smart farming, and insurance risk. The data is a critical enabler for everyone from researchers to multinational corporations.
Along with unprecedented opportunities, the sheer variety of geospatial data, and ways to analyze it through artificial intelligence, pose a challenge. Even the biggest players struggle to locate the best data sources and AI algorithms scattered around the world. For small teams, hurdles can seem insurmountable.
“Wherever data sets may be located in far-flung parts of the world, we're able to crunch it for our clients at planet scale. This simply would not be possible without Google Kubernetes Engine and Google Cloud data centers in every corner of the world.”
—Barry Nagel, CTO, UP42This challenge led to the founding of UP42, a platform that provides a one-stop shop for anybody with a need for geospatial insights to find the right data and algorithms for them. The platform then enables clients to build their own data solutions with world-class processing capability and an intuitive dashboard that gives them instant insights.
To succeed, UP42 needed virtually unlimited computing power and versatility to run CPUs and GPUs at planet scale, anywhere geospatial data sources may be hosted. The team found the answer in a combination of Google Kubernetes Engine (GKE) and the Google Cloud global network of data centers, which provided the unlimited scalability to meet client needs in real time.
“We bring access to Earth observation data and processing power together. Wherever datasets may be located in far-flung parts of the world, we're able to crunch it for our clients at planet scale,” says Barry, CTO at UP42. “This simply would not be possible without Google Kubernetes Engine and Google Cloud data centers in every corner of the world.”
“Google Kubernetes Engine lies at the very core of our mission. It enables an endless variety of complex tasks, from uploading massive amounts of geospatial data to spinning GPUs for AI algorithms. That would be impossible with a less powerful or flexible infrastructure.”
—Barry Nagel, CTO, UP42Enabling a lofty vision that brings geospatial data down to earth
In “The Hitchhiker's Guide to the Galaxy,” the number 42 offers the answer to “life, the universe, and everything,” says Barry. “It's the vision that inspired us to put 42 in the company's name.” (UP, he says, refers to “devices that soar above us, satellites, drones and planes, that give us life-enabling Earth observation data.”)
In other words, the UP42 mission is to literally bring stratospheric insights down to earth. “Our goal is to lower the barriers to entry for geospatial data analysis,” says Barry. “That means enabling innovators small and big to instantly access Earth observation data from our library of sources and model it with AI algorithms, all driven by the unlimited processing power of Google Cloud.”
Google Cloud's global network of data centers represents the simplest and most obvious way UP42 brings processing and data together, enabling low-latency geospatial data uploads anywhere in the world. The second critical driver is GKE, which brings versatile, unlimited scale to all of UP42's core operations, says Barry. “We have several dedicated Kubernetes clusters running simultaneously to power every key dimension of our platform,” he says.
First UP42 deploys a Kubernetes cluster to run services and products from frontend user interface (UI) to backend data retrieval. Next, the platform provisions multiple dedicated clusters for uploading massive geospatial payloads in different world regions.
Lastly, it runs a variety of specially configured clusters for processing the algorithms that draw intelligence from the data. These range from geospatial imaging that requires only CPU and memory, to AI-based algorithms that need GPU for more complex tasks such as object detection.
“Google Kubernetes Engine,” says Barry, “lies at the very core of our mission. It enables an endless variety of complex tasks, from uploading massive amounts of geospatial data to spinning GPUs for AI algorithms. That would be impossible with a less powerful or flexible infrastructure.”
“We simply distributed all of our processing needs onto Google Cloud, ramped up the infrastructure as much as needed, and delivered super-fast execution. We were still fresh in the market, and our client told us, ‘we've been testing many platforms. You're the only one that's able to deliver this scale.’”
—Barry Nagel, CTO, UP42Impressing clients with instant terabyte-level data processing
The sheer power of GKE became clear to Barry from the early days of UP42. A customer asked the fledgling platform to perform a proof-of-concept for oil leak detection that included synthetic aperture radar (SAR), a data-intensive radar technology that generates high-resolution remote sensing imagery.
Thanks to Google Kubernetes Engine, UP42 was able to execute 350 geospatial processing jobs, ramping up to 870 virtual cores and crunching six terabytes of SAR data, within one hour. “This was amazing even to us,” Barry recalls.
“We simply distributed all of our processing needs onto Google Cloud, ramped up the infrastructure as much as we needed, and delivered super-fast execution,” he says. “We were still fresh in the market, and our client told us, ‘we've been testing many platforms. You're the only one that's able to deliver this scale.’”
Enabling data insights and partnerships with Looker
UP42 needs more than data crunching power to succeed. Its future depends on enabling accessible, intuitive intelligence for clients and partners from the masses of satellite data it feeds through GKE. Powerful enablers of that mission are BigQuery and Looker.
For Barry, one of the most important use cases of the data analytics and dashboard platform is the way it enables partnerships with owners of datasets and AI algorithms, who contribute to UP42's library of solutions.
In any given project, a customer may select a data source owned by one company, a processing algorithm built by UP42, and an AI algorithm created by yet another company. That means a single client workflow may deploy the intellectual property (IP) of multiple different companies, which then must share the revenue.
UP42 needs to keep track of ownership of every block of data that runs through the workflow. “We always need to know how many jobs have been executed, with whose IP, down to the last pixel or square kilometer,” explains Barry. “Looker plays a big role in enabling the revenue sharing. We've built Looker dashboards specifically for our partners to track the breakdown of data usage based on IP.”
Drawing creative inspiration from being Google Cloud native
UP42 began life on Google Cloud and its entire architecture is hosted on it. Being Google Cloud native, says Barry, has sparked unexpected inspirations for building a solution that removes barriers of entry to innovating with Earth observation data.
From the start, Barry says, the team felt that Google Cloud was simply more intuitive and easier to use than other providers. And as UP42 began developing its own user-friendly geospatial solution, it sought clues from the Google Cloud console UI, which enables developers to create without complications.
“Being Google Cloud native gave us insights into how to make our own geospatial data solution smooth and easy for our customers,” says Barry. “The Google Cloud ecosystem, especially the dashboards and UI, are extremely easy to use. That's a model we wanted to emulate for our own clients. And it's a key reason for our success.”
Building the geospatial future with Anthos and TensorFlow
Barry envisions an unlimited future with Google Cloud as geospatial data gains importance in solving global challenges, from climate change to water shortages.
UP42 currently deploys TensorFlow for its own AI modeling needs as it bolsters its library of geospatial algorithms. The next step will be to enable clients themselves to train and build ML models on the UP42 platform. TensorFlow promises to play a big role.
Barry says another future UP42 project is to expand its potential client base to hybrid cloud or private cloud users. To that end, UP42 sees Anthos as the way forward. Barry says it will “enable different processing clusters on different clouds, in different environments, that then automatically connect to our platform.”
“We want to support more use cases and more types of companies that have very specific needs,” says Barry. “Anthos can be a big driver for this. It's a powerful example of our exciting and open-ended relationship with Google Cloud.”
Tell us your challenge. We're here to help.
Contact us
About UP42
UP42 is a developer platform and marketplace for Earth observation data and analytics that help to solve problems ranging from glacier erosion to oil leak detection. The platform offers a library of geospatial data sources and AI algorithms enabling clients to build their own solutions using UP42's API and unlimited processing power. UP42 is a subsidiary of Airbus SE.