Solution Architect Matthieu Vansteene played a key role as the lead architect for Capgemini in the Domino-X project. He worked on defining the interfaces and architecture of the Domino framework for Earth Observation (EO) ground segments.
Now, as part of the Domino-E project, Matthieu is building on that foundation by developing the first practical Dominoes, turning the innovative concepts of Domino-X into operational reality. In this interview, he discusses the evolution of the Domino architecture, the challenges and benefits of its cloud-based design, and its potential to revolutionize how EO ground segments are developed and operated.
The Domino-X project set the framework for the Domino architecture, and now Domino-E is one of the projects that actually begandeveloping Dominos. Is that correct?
As you say, in Domino-X we defined the framework, and now via Domino-E, we start to build three Dominos. One for the Coverage Service, which is part of the federation Domino. One for SCRMS, which is an algorithm for antenna selection, either antenna as a service or private-owned antenna. And there is a third Domino concerning a virtual assistant to allow an end-user to order an acquisition, which is part of the user access service dominoes.
From your professional perspective, as a software engineer, what’s the big novelty, the big advantage of the Domino architecture?
In Domino, we focus on several topics. First, it’s a cloud-native architecture by design. We really focus on facilitating deployments on the cloud, whether that is a public cloud or on-premise cloud. Which represents an innovation in this field. The second one is standardization. Currently, there is no well-established standard for building ground segments. And so, via Domino, we have defined a standard for the future.
And the third part is the Domino structure of the project. Each building block is a Domino with clearly defined responsibility and interfaces. You can build your ground segment by gathering Dominoes from a catalog. You can choose a Domino, adapt a Domino or even build your own Domino if the catalog doesn’t fit your needs.
That’s the big principle of Domino-X, modality.
So it’s like you created standards and interfaces, and it’s like sort of an operating system. And now people can develop their own software packages, let’s say, for it. Is that an analogy that would be describing it in a way, or is that totally wrong?
Yes, that’s really the principle behind Domino. Where interfaces are defined, but within each Domino it is a black box for the user. Each company can build what they want inside. They are responsible for the functionalities, although the Dominoes must adhere to certain specifications. The Domino-maker is responsible for performance, but they have the freedom to develop what they want. So they can develop new software, or they can take legacy software, already flight proven, put it in the Domino, adapt the interface, and add it to the Domino catalogue.
There is another aspect, too, in the form of test Dominoes. If you have a new concept, for example a new algorithm, you can develop a new Domino and deploy it in parallel to an existing Domino to test it. It’s really a game of Dominos.
To what extent does this change the way the ground segments were developed before?
In fact, there is no standard for ground segments, so from one ground segment to another, it’s very difficult to reuse components, so companies tend to develop from scratch. This changes with Dominoes as each Domino is interchangeable and so you can reuse them from one mission to another. And from one company to another, because you can decide to take a Domino from a company and after that to change with another equivalent Domino from a different company. And for me, that’s a big new way to build a ground segment.
Do companies have open access to the Domino standards, Interfaces and protocols?
Yes, all those documents, including ICDs and models, are freely available on the Domino-X website, which is https://domino-x.space.
Does the development of Dominos for Earth observation ground segments require a lot of expertise and maturity in the space domain?
That depends greatly on the Domino. For example, a mission-specific Domino would need a very fine knowledge of the spacecraft structure. But if you are interested in image processing, you can develop an algorithm fairly easily, put it in a Domino, and add it to the catalogue either commercially or as open source. So, Domino is more orientated towards industry, but remains accessible to researchers or students.
Apart from the three dominoes from the Domino-E project, are there any other examples of components that are currently under development?
There are Dominos that cover the full ground segment, from mission control instructing the spacecraft to capture images right to the final stages where images are received, processed, and distributed. Additionally, we have new Dominos like the prototype “Decision Maker.” This Domino collects alerts from sources such as social media, detecting keywords like “flood”, so that when a flood is identified, it can automatically trigger image acquisition for fast response. Dominos encompass traditional ground segment functions and new capabilities like these.
You’re working on Domino projects for 4 years now. In the beginning, when you began working on Domino–X back then, what did you believe was fascinatingabout the approach?
For me, it was my first experience in space mission control. Over time, I’ve had the opportunity to really contribute to something new. The Domino concept evolved gradually during the 2-3 years of the Domino-X project, which was both a new and fascinating experience. The project involved many major companies, including Airbus Defence & Space, Thales Alenia Space, Safran, CS Group, and Capgemini. It was an opportunity to collaborate, share expertise on satellite and ground segments, and gain broad knowledge in the space domain.
And what are you working on at the moment?
I’m currently responsible for designing the testbed in a cloud environment. This decision aligns with the goals of Domino X and addresses the fact that we are working with many European companies spread across different locations. By using the cloud, we ensure easy access for all parties involved. It’s also a practical solution—if the testbed were hosted by a single company, it could create access challenges due to security concerns. Hosting it in the cloud avoids these issues and simplifies collaboration.
How will the test bed be used?
At Capgemini, we will implement the infrastructure, including all the testing tools. Each company will then create scenarios, build simulators, define KPIs, and set up logs. They can run their scenarios on the testbed, collect the results and analyze them. It’s a two-part process: Capgemini provides the testing capability, and the companies conduct and analyze the tests.
What’s a typical test scenario?
For example, with the Coverage Service Domino, we define a large area on a map to cover, and the Domino splits it into smaller segments for acquisition. We can then simulate acquisition failures, such as cloud cover affecting image quality, prompting a reassessment. At the end, we collect the final product and evaluate how long it took to achieve 90% or 100% coverage of the area. This allows us to adjust the algorithm, test its impact, and optimize the process.
When you take a personal look into the future of the Domino architecture, what do you see?
First, I hope that the Domino concept will be widely adopted. Some Dominos are critical to our mission, while others are optional or offer additional services. I believe we will begin integrating Dominos gradually, starting with a mix of the old system and the new Domino architecture, and then transition step by step to a full Domino-based system. There’s a lot of work ahead—first, to implement the core Dominos that form the foundation of our system, and then to add new Dominos for additional functionalities.
Thank you!