New director for the Institute of Software Technology, Professor Dr. Michael Felderer

His predecessor, Rolf Hempel, retired at the end of 2022 after a good 20 years of management. He is handing over to the new director an institute that now has around 140 employees at seven DLR locations in Germany and deals with research topics ranging from software engineering and high-performance computing to scientific visualisation and quantum computing.

In an interview, Prof. Felderer explains how his career as a software engineering expert led him to the German Aerospace Centre. He also takes a look at current and future topics at a software institute in the midst of a rapidly changing software landscape.

Mr Felderer, over the past five years you have held professorships in software engineering at the Universities of Innsbruck and Blekinge in Sweden. You have also worked as a managing director in software consulting. What attracted you to the idea of the German Aerospace Centre as your next stop?

The German Aerospace Centre offers a unique research and innovation environment in the fields of aerospace, aviation, energy, transport and security. These are subject areas in which DLR has a great deal of internationally leading expertise and which fascinate me personally. In space, aeronautics, energy, transport and security, the potential to realise new product innovations and services or make existing ones more efficient and effective with the help of software is far from exhausted. Just think of the topic of New Space. However, this can only be achieved on the basis of scientifically sound software engineering methods.

Which topics do you see as particularly relevant for the development of competitive modern software in the next five to ten years? Do you also see challenges ahead in these areas?

An important trend is the provision of software as a service, for example in cloud environments. This provision of software-based services is also becoming increasingly important for engineering applications, for example for digital twins, both inside and outside DLR. This transformation from standalone software to services from the cloud offers a lot of potential, especially in transfer projects, but also leads to new software-related challenges that have not yet been analysed in detail. On the one hand, quality characteristics such as security, performance and reliability are becoming even more important. On the other hand, it also requires software development approaches that enable the appropriate structuring of software into independently deliverable units, so-called containers, and their orchestration. Especially for the provision of digital twins from DLR's application areas as services, there is still a lot to investigate and implement in reliable software-based solutions.

Generative approaches to artificial intelligence (keyword: ChatGPT) require new approaches to quality assurance in order to be able to check the generated artefacts, such as text or code, for their robustness, fairness or other quality characteristics. These approaches will also change software development itself: Traditional manual software development will be complemented by AI-supported digital assistants that need to be developed and whose interaction with software developers needs to be investigated and utilised for the DLR and beyond.

New approaches to visualisation, high-performance computing and quantum computing enable even more powerful simulations and digital twins with high data resolution, which can be used for new types of material simulations or in virtual product certification, for example. The interaction of security and safety also plays an important role, especially for product certification, in order to be able to provide resilient products. The development of corresponding automated analysis methods that integrate security and safety is still in its infancy, but is essential for the use of systems in the cyber age.

What synergy effects do you see in combining the management position of a research institute with a university professorship?

The management of a research institute such as the DLR Institute of Software Technology and an associated university professorship are mutually beneficial. On the one hand, research at a large research organisation like DLR offers opportunities that are not normally available at a university. In my field of research, for example, this concerns the development of innovative quality assurance approaches that can be tested in the context of space systems or quantum computing, but also the use of powerful simulators in aviation with high data resolution. These possibilities are hardly available at a university. However, the findings of non-university research at DLR also provide new impetus for university research, the results of which, especially in the form of new methods, then enrich non-university research. One example of this could be test case generation based on AI algorithms, which is being further developed at the university and applied to powerful simulations at DLR in order to efficiently identify safety-critical scenarios, for example.

In addition, combining the management of a DLR institute with a university professorship is also a good opportunity for cooperation with other colleagues at the university and for joint research projects between DLR and the university. Being anchored in university teaching also offers the opportunity to get students interested in our topics at DLR and to integrate them into our projects as employees.