Innovation is in Audi’s DNA.

So, it’s no surprise that after more than a hundred years of automotive excellence, the company is pushing new boundaries at its Böllinger Höfe plant.

It’s latest goal? A fully software-defined smart factory.

This month, Audi’s plant, which is in Germany’s Baden-Württemberg region, announced the initial rollout of the first virtual programmable logic controller (vPLC). This is part of a broader strategy to bring IT into the OT domain, and it’s driven by Audi’s new Edge Cloud 4 Production (EC4P) solution.

Developed with support from key partners like Cisco, EC4P enables the next level of virtualized, smart automation. So far, it is managing automation cells for the car-body assembly line at the Böllinger Höfe factory. But moving forward, it promises highly efficient control of even the most extensive arrays of robots and systems, while reducing hardware, lessening energy needs, and increasing security.

To learn more, we spoke with Sven Müller, Audi’s project manager for Edge Cloud 4 Production, and Thomas Kampa, IT solution architect.

Thanks for meeting with us today, Sven and Thomas! Sven, we’ll start with you. Audi is a longtime leader in industrial IoT and smart-factory automation. Where do you stand today?

Sven: Thank you, Kevin! From the production side, we have greatly improved our processes in recent years. We’ve excelled in optimizing the hardware and processes — for example, the robotics and systems that help to build the cars. But now we’re taking the next step, toward a fully software-defined factory. That’s our target.

That brings us to Audi’s Edge Cloud 4 Production initiative. What challenges was it designed to meet?

Sven: The EC4P project is a virtualized platform that introduces a totally new idea in the manufacturing domain. We had a rapidly growing number of industrial PCs in our production and now we are shifting paradigms. Automation is a lot about hardware today, and we want to rethink that. We want to focus now on applications, and EC4P is a platform on which all of the production-related applications find a new place to run. That also means that now applications are communicating with our robots through the network in real time, which nobody has done until now.

So, it’s cloud based, it’s platform based, and it’s virtualizing pretty much every aspect that can be virtualized.

Sven: Right, it’s cloud based, but it’s an on-premise private cloud within our data centers. For certain use cases, latencies in the microsecond range have to be satisfied. That can only be fulfilled with a private cloud on our own ground, and a powerful new network that can transport all the packets, all the data, back and forth between the data center and the plant floor. For instance, to fulfill the industrial standards for communication that our robots are setting, we need less than one millisecond of end-to-end latency.

Audi’s a high-profile company, so obviously security is a key concern as well.

Sven: Yes, and even with a private cloud, if you have a big network with thousands of assets in it, it’s still potentially at high risk. One tool to mitigate the chance and impact of cyberattacks is micro segmentation. Cisco’s software-defined access (SDA) solution provides a fantastic way to manage and police the communication between our assets, physical or virtual.

Thomas, as an IT solution architect you’re no stranger to uses cases demanding microsecond latencies. Tell us more about this month’s milestone.

Thomas: So, the highlight here is that we are bringing online the first virtual PLCs — that’s programable logic controllers. The PLCs are the brain and heart of every factory that control the manufacturing processes, including the robots. But at the same time, they allow us to safely operate our automation devices. Normally, you have this as hardware directly next to the robot, but now we host virtual safety PLCs in our data center. And the second thing that I want to highlight is the new network. Highly critical communication like this requires a highly reliable, deterministic network to get from the data center to the respective robot cell. And so, together with Cisco we have created a dual-fabric concept enabled by SDA. If one fails, we have the second fabric as a backup. It is of the utmost importance that robots are up and running and producing beautiful cars.

What other Cisco technologies are supporting that reliability?

Thomas: We mentioned the micro segmentation and the virtual PLC use case. We’re gradually adding micro segmentation to further parts of the production plant to protect applications like our worker guidance system. Here, we are also leveraging Cisco’s Identity Service Engine. It comes with the fabric, and it’s giving us more visibility into what devices are communicating in our network.  Also, in terms of increasing our visibility, we’re exploring Cisco’s Splunk and Cisco’s Provider Connectivity Assurance solution. Because at the end of the day, we want to have end-to-end visibility from the PLC in the data center down to our robot cells. We want to see what’s going on across the platform, what’s going on in the application, and what’s going on in the network. We’re creating solutions together with Cisco because it’s crucial to our operations team that we have the best visibility solutions and that we improve processes and designs based on data-driven decision making.

AI is already integrated into many of these Cisco solutions, but what role will it play moving forward?

Sven: AI will play an increasing role in our next steps. For example, as we bring the full stack into production, it will put a heavy burden on our maintenance guys for configuring all this stuff. That is one place where AI can help simplify things. We’re also exploring how we can use computer vision to drive better decision making and quality control.

Sustainability is a key concern for Audi. How is EC4P — and the network — helping?

Thomas: Starting with EC4P, we are replacing thousands of full-blown industrial PCs with commodity servers. So, we are gaining energy efficiency and reducing our carbon footprint through virtualization. This is only possible with a reliable and fast network in place.

How can Audi’s innovations influence the automobile industry and others in the manufacturing space moving forward?

Sven: This idea of the EC4P stack — and the whole architecture we are talking about, from the application, the platform, the network — is something which we are not seeing in the manufacturing industry today. So, we see our role as leading the industry to this transformation.

What else do you see for the future?

Sven: The EC4P project is the foundation for the Audi factory shopfloor of the future. With that in place, we can start leveraging the full potential of software-defined automation and AI in our factories

And Cisco is helping provide the foundation for that future.

Thomas: Yes, it’s crucial to have a high-performance, resilient, and secure network. And it’s a strong partnership. We have a lot of ideas that we bring to Cisco, and then we develop them together and integrate them into a working solution. New concepts and features are tested rigorously in international testbeds, especially together with Cisco’s IOT unit. The community and the collaboration between the teams is a big part of our success. Together, we’re driving a new paradigm of manufacturing.