Data Ecosystem and Asset Administration Shell

Why am i writing this article?

If you’ve been following trends in the German manufacturing industry, you’ve probably heard a lot about Data Ecosystems and Asset Administration Shells (AAS). These two topics are gaining serious traction — so much so that the German government is pouring significant money and resources into them.

But here’s the thing: while I’ve seen some of articles diving into these concepts individually, I haven’t come across much that explains how they connect — and why that connection matters. The convergence of Data Ecosystems and AAS is important, especially as they shape the future of how industries manage data and digital twins.

So, after being involved in projects related to these topics for the past few years, I thought, why not help shed some light? If you’re trying to get your head around these concepts, I’m hoping this will help clarify things a bit.

Who should read this article?

• If you’re new to the idea of Data Ecosystems and want a clearer understanding, you’ll get some insight here.
• If you’ve been wondering what Asset Administration Shell have to do with Data Ecosystems, this is for you.

Now, I’ll be honest — this article simplifies a few things, so if you’re already an expert, you might find it a bit light. But for those just getting started, it should be a helpful guide.

(Note: I’ve shared all the relevant links in the first comment of this article, so you won’t get sidetracked while reading.)

What is Asset Administration Shell?

Let’s break this down simply: the Asset Administration Shell, or AAS for short, is like a rulebook that sets the standard for how we describe something called Digital Twins. And what exactly is a Digital Twin? Think of it as a digital copy or mirror of a real-world object, process, organization, person — basically anything you can think of, even abstract concepts.

To make it even clearer, imagine a robot working on a factory floor. Using the AAS standard, we would create a Digital Twin of this robot, which would be represented by something called an Asset Administration Shell object. The core thing here is that this Digital Twin needs a globally unique identifier (kind of like its digital passport). Beyond that, it can also include things like a display name, a description, or other bits of info that add more context.

Now, this robot has all sorts of info tied to it: technical specs, user manuals, real-time operational data, etc. In the AAS world, all these different pieces of information are broken down into something called Submodels. These Submodels let you view the Digital Twin from different angles, each giving specific details — whether it’s how the robot operates, how it was built, or how it’s maintained.

If you’re curious to dive deeper into what AAS is all about, I recommend checking out my other article, “Journey into the World of Asset Administration Shell.”

What is a Data Ecosystem?

Alright, let’s get to what a Data Ecosystem is. Think of it like a network where different companies (the data producers, consumers, and anyone in between) come together to share, analyze, and use data. Sometimes, people refer to these ecosystems as DataSpaces, but for now, let’s treat both terms as the same thing. The key idea is that this collaboration makes it easier for everyone involved to make better decisions and work more efficiently by using shared data.

Let me give you an example to make it clearer:

Imagine there’s a Robot manufacturing company called Alpha Robotics that builds industrial robots. They get their electric motors from another company called Beta Electric. Now, Beta Electric, in turn, gets the bearings for those motors from two other companies: Gamma Bearing Company and Delta Bearing Company.

Here’s where things get interesting. Alpha Robotics needs to report the CO2 footprint of their production to regulatory authorities. The CO2 footprint isn’t just about what they produce; it also includes the CO2 footprint of the parts that come from Beta Electric and their suppliers. So, Alpha asks Beta Electric for their CO2 footprint. Beta Electric can’t just give their own number — they need to add up their own production CO2, plus the CO2 from the bearings they get from Gamma and Delta.

Sounds simple enough, right? Now imagine trying to automate this whole process so that data flows smoothly from Delta and Gamma all the way up to Alpha. In the real world, this is a huge challenge for a lot of reasons — different systems, privacy concerns, data quality, and so on.

This is where the Data Ecosystem comes in. It’s an attempt to make this whole data-sharing process easier and more efficient, allowing companies to seamlessly exchange information like CO2 footprints (or any other data) across a network.

The Challenges of Data Sharing

Now, let’s talk about the real-world problems of getting data to flow smoothly between companies. As much as we’d like it to be easy, it’s actually pretty difficult, and here’s why:

1. Different Data Formats Across Companies

Think back to the earlier example with Gamma and Delta Bearing Companies. They’re trying to share the same kind of data (like CO2 footprint), but here’s the catch: each company stores that data differently. The names they use for certain properties, the descriptions, and even the units of measurement might not match up. One company might use “kg of CO2,” while another uses “tons of CO2.”

Now, imagine Beta Electric buys bearings from 20 different suppliers. If every supplier uses a different format or way to label their data, Beta Electric is stuck with the messy job of trying to compare, consolidate, and make sense of all that info before they can integrate it into their system. It’s a data nightmare!

2. Different Technical Infrastructures

Even if they figure out the data format issue, another challenge is the way each company shares that data. Some suppliers might still send printed paper documents (yes, that still happens!), others might email a PDF or an Excel spreadsheet, and the more advanced ones might have an API (Application Programming Interface) that allows Beta Electric to automatically pull the data into their system.

Here’s the problem: if Beta Electric is working with 10 different bearing manufacturers and each one uses a different method to share data, automating this process becomes nearly impossible. They’d need to create a unique integration for each supplier’s system, which is complex and time-consuming.

Because of these two big challenges — different data formats and incompatible technical infrastructures — getting data to flow seamlessly across companies is, to put it mildly, a logistical headache. It’s no wonder that smooth, automated data sharing between multiple companies is still close to impossible in many industries.

How Data Ecosystems Help?

So, how do Data Ecosystems solve the challenges we just talked about? They tackle the two main problems by:

1. Standardizing the Data Exchange Format

Data Ecosystems enforce the use of industry standards, like the Asset Administration Shell (AAS), to make sure everyone’s on the same page when it comes to data formats.

What does that mean? Let’s revisit our example: Gamma and Delta Bearing Companies need to share data with Beta Electric. Instead of sending the info in all sorts of formats, they both use the same, consistent format (same set of attributes, same set of field names, same unit of measures etc) thanks to AAS. This makes it so much easier for Beta Electric to integrate the data without having to decode it first.

2. Unified Technical Infrastructure for Data Sharing

The second big win is creating a unified technical infrastructure that companies can use to share data. If you want to exchange data within a Data Ecosystem, you need to follow specific architecture guidelines, including tools like the AAS Digital Twin Registry and the Data Space Connector. There are some other technical components involved, like a decentralized wallet, but let’s keep it simple for now.

What is the AAS Digital Twin Registry?

Think of the Digital Twin Registry like a phonebook for Digital Twins based on the Asset Administration Shell standard. It doesn’t store the actual data but acts as a directory, showing where different Submodels (those smaller chunks of data we talked about earlier) are located. This makes it easy to retrieve the data via APIs (automated connections between systems).

What is the Data Space Connector?

The Data Space Connector comes from a concept called Data Spaces (governed by the IDSA — International Data Spaces Association). There’s an implementation of this called the Eclipse Data Space Connector, which is used in Data Ecosystems like Catena-X. Without diving too deep into the tech side (like Data Plane, Control Plane, etc.), let’s focus on what the Data Space Connector actually does.

Simply put, the Data Space Connector acts as a gatekeeper for the data that comes in or leaves your company’s boundaries. It helps with two things:

1. Negotiating Data Access: It allows companies to negotiate which data should be shared, with whom, and under what conditions.
2. Data Exchange: Once the negotiation is done, it securely funnels the data request from the company that needs the data to the company providing it.

How Does This Work in Practice?

Let’s go back to Alpha Robotics, which buys electric motors from Beta Electric. Alpha Robotics needs the CO2 data, along with some technical info, from Beta Electric. Here’s how it works:

• Alpha Robotics’ Data Space Connector reaches out to Beta Electric’s Data Space Connector and starts negotiating. For example, Alpha might say, “We’d like access to CO2 data with fields A, B, and C for the next 6 months.”
• If Beta Electric agrees, its Data Space Connector opens a secure channel with Alpha Robotics.
• Alpha Robotics’ Data Space Connector then sends a request for the data to Beta Electric’s Data Space Connector.
• Beta’s Connector checks with its Digital Twin Registry to find out where the data is stored.
• Once it knows where the data is, it sends that address back to Alpha Robotics so they can fetch the real data.

And just like that, the whole process of sharing complex data across company boundaries becomes a lot smoother and more efficient.

Is Everything Really So Straightforward?

If you just read the example I gave above, you might think this all sounds pretty straightforward. But having worked closely with this for a few years, I can tell you there are some real-world challenges that make things a bit trickier.

The Chicken-and-Egg Problem

Let’s say Beta Electric wants to join a data ecosystem. The catch? They’ll only really benefit if all their bearing suppliers and Alpha Robotics are also part of the ecosystem. So Beta Electric might decide to wait until their suppliers and Alpha Robotics join first. But guess what? Alpha Robotics is probably thinking the same thing — they’re waiting for Beta Electric to join. And the suppliers? Well, they’re likely waiting for Beta Electric too. It’s a classic chicken-and-egg situation, and it makes adoption of data ecosystems a slow, hesitant process.

Academia Meets Business

Data Spaces and concepts like this were developed in academia for a long time. But when these academic ideas meet the practical, day-to-day realities of business, unexpected problems often arise. Solving those problems requires rethinking things from a conceptual, technical, and business perspective. And in such a complex and highly distributed environment, this takes time and effort — sometimes a lot of both. I won’t dive into the details here because it could make this article much longer, but if you’re curious, feel free to reach out.

Investment Challenges

As I mentioned earlier, joining a data ecosystem means you’ll need to adopt several technical components that all need to work together smoothly. Now, if you’re a small company without a dedicated IT team — just business users — it’s going to be a challenge. You’ll either need to get familiar with all the technical ins and outs, or you’ll have to invest in resources to manage it for you. And beyond the workforce, there’s also the financial investment in the technology itself.

Naturally, companies will only make this investment if they see a clear return on it. And we’re back to the chicken-and-egg problem again: you’ll invest only if others join, but they’re waiting on you to invest first.

The Optimistic View

Now, don’t get me wrong — I’m not a pessimist. I’d say I’m more of a rational optimist. While the challenges I mentioned are real, there are a couple of reasons why I believe we can overcome them.

1. The Role of Government Investment

One of the biggest reasons for optimism is the huge investment the German government is making in industry-specific data ecosystems. Here are a few examples of ecosystems that are tailored to specific sectors:

• Catena-X for the automotive industry: A collaborative platform aimed at improving efficiency and innovation across the automotive supply chain by standardizing data exchange.
• Factory-X for manufacturing: Focused on smart manufacturing and Industry 4.0, helping companies use data to automate processes and make better decisions.
• Aerospace-X for aerospace: Makes data sharing within the aerospace industry easier, improving operational efficiency and innovation.
• Process-X for process industries like chemicals and pharmaceuticals: Helps with data integration and analytics, ensuring regulatory compliance and improving quality.
• Semiconductor-X for the semiconductor industry: Tackles the data-sharing challenges unique to the semiconductor sector, making the supply chain more efficient.

These industry-specific ecosystems are designed to solve many of the problems I talked about earlier. For example:

• The Chicken-and-Egg Problem: With platforms like Catena-X, if all automotive manufacturers and their suppliers are required to use a standardized data exchange format, it could trigger mass adoption. No more waiting around for others to join first!
• Academia Meets Business: These initiatives involve both companies and research organizations, working side by side to optimize technical solutions based on real business needs. This collaborative approach is helping to bridge the gap between theoretical concepts and practical use cases.
• Investment: As these initiatives grow, they attract more software vendors, which means smaller companies will have access to easier, out-of-the-box solutions. This lowers the barrier to entry for companies that want to join the ecosystem but don’t have huge IT budgets.

2. The Data Compliance Wave in the EU

Another reason to be optimistic is the wave of data compliance regulations coming out of the EU. Initiatives like the EU Digital Product Passports and Extended Producer Responsibility are forcing companies to consolidate and share data with external parties, whether it’s with end consumers or regulatory bodies.

The bottom line? Companies will have to do this anyway, because it’s a matter of compliance. But if they’re smart about it, they can align these regulatory requirements with their participation in data ecosystems, which will boost their return on investment.

Conclusion

The journey towards seamless data exchange through Data Ecosystems and the Asset Administration Shell is far from straightforward, but the potential benefits make it worth the effort. With growing government investment, industry-wide initiatives, and the push for compliance, we’re seeing real progress in overcoming these challenges. Those willing to adapt and embrace these new standards will be better positioned to innovate, collaborate, and stay competitive in an increasingly data-driven world.

Let’s Connect

I’d love to hear your thoughts on these topics. Whether you have feedback, questions, or ideas on how to navigate the challenges of Data Ecosystems, feel free to reach out. Collaboration is key in shaping this evolving space, and I’m always open to learning from others. Let’s keep the conversation going!