Software and data expert Prof. Bernhard Mitschang brings together computer science and production. Their almost complete networking and comprehensive digital data production and provision can be found in the digital real-time capable factories, the so-called advanced manufacturing, in the entire field of Industry 4.0 and also in the cyber-physical IT systems. Trending terms in this area are Big Data, Data Mining, Machine Learning and also Data Science. Mitschang and his team have adapted their research to best answer the questions posed by Big Data and data analysis in the respective industrial application context.
The traditional areas in which databases and information systems were used just 20 years ago included banks and insurance companies, the business world, technology and science. “The Internet,” as Bernhard Mitschang, who has held the Chair of Databases and Information Systems at the University of Stuttgart’s Institute for Parallel and Distributed Systems since 1998, remembers, “still didn’t exist in its current form back then; apps were a foreign word, and the technology was still in the Stone Age.” The turning point came with the Internet of Things (IoT). “The Internet, and especially the IoT have brought about a radical convergence between manufacturing and information technology,” Mitschang explains adding that this is reflected in an almost complete interconnectivity and a comprehensive digital data production and provision system. “The many social media applications are examples of this as are the digital, real-time factory, so-called ‘advanced manufacturing technology,’ everything that falls within the sphere of Industry 4.0, and cyber-physical IT systems."
Mitschang himself played a role in this development at an early stage. Shortly after his appointment at the University of Stuttgart in 1998, he contacted Prof. Engelbert Westkämper, the then spokesman for the Collaborative Research Center SFB 467 (Adaptable Corporate Structures for Highly Variable Series Production) and one of the pioneers in the field of advanced manufacturing. “I’ve been collaborating with mechanical engineers ever since,” says Mitschang. He was also involved in the “Graduate School of Excellence Advanced Manufacturing Engineering” (GSaME), one of the University of Stuttgart’s first two excellence projects, right from the start and has been its spokesperson since 2014.
The Internet still didn't exist in its current form back then; apps were a foreign word and the technology were still in the Stone Age.
Prof. Bernhard Mitschang
"Data science" degree program launched
The Chair’s research focus has changed in response to the digital transformation in industry. “We are currently working on issues that have to do with data provision, data management and data analysis with a view to developing and supporting the applications of digitalization. The relevant buzzwords include Big Data, Data Mining, Machine Learning and also Data Science.” Mitschang and his team are taking a twin-tracked approach to tackling the challenges that underlie these buzzwords. On the one hand, the Chair has adapted the range of degree courses on offer and established the “Data Science”, bachelor’s degree program – a rarity in Germany. On the other, the research focus has been adapted to enable researchers to provide the best possible answers to the questions of Big Data and data analysis in their respective industrial application contexts. The “ICT Platform for Production” junior research group at GSaME, which has been headed up by Dr. Peter Reimann since 2017, is one example of this. The group focuses on an information and communication solution that not only integrates production processes, but also the heterogeneous, distributed information systems throughout the company as well as such things as mobile devices. The projects are defined in collaboration with industry partners such as Daimler, Festo, Trumpf or Mann+Hummel and supported via the individualized qualification program. They are then implemented in compliance with the GSaME’s quality standards. “We are extremely careful to ensure that rather than simply being chosen and implemented on a ‘let’s do it’ basis, all work carried out is scientifically relevant,” Reimann emphasizes.
Ict platform for the entire product life cycle
At an early stage, the group developed the “Stuttgart IT Architecture for Manufacturing” (SITAM), which enables companies to acquire, manage and analyze data. The data analysis functions in the SITAM are currently being expanded. Implementing data analysis technologies across the entire product lifecycle rather than just in certain individual production phases is the aim of industrial analytics. “This gives us a better understanding of and enables us to optimize such things as products, entire factories and individual machines,” as Reimann explains.
Making sense of the data lake
The basis for such holistic analyses are so-called data lakes, which are highly scalable databases into which the raw data generated along the value creation chain flows, making data lakes a very flexible basis for data analysis. However, as Doctoral Researcher Rebecca Eichler explains, the problem is that “companies often don’t even know what data their ‘data lakes’ contain, as it is inadequately described.” Among other things, the systematic storage and management of the enormous data volumes to create added value requires one to document what data exists, what it describes, its quality and origin, and who is permitted to access it. So, data management relies on so-called metadata, i.e., data about the data. It is precisely this metadata management to which Eichler is devoting her doctoral project “MetaMan” (under the supervision of Dr. Holger Schwarz), which she and Bosch are working on together. The relevant question goes beyond data lakes to take in the entire corporation. “Until now,” she explains, “metadata management has been focused on individual sub-processes or corporate departments. What we are trying to do in our project is to develop techniques and concepts for designing a metadata management system for the entire corporate structure so that, for example, data can be made available across departments.” This is inspired by existing inter-company data marketplaces. One can think of them as platforms, which use metadata to enable users to find, understand and access data as well as to upload their own data. Very often when this concept is transferred to the internal corporate environment, more sensitive data is traded on the data market. “This means that issues of transparency and compliance take on a whole different level of significance,” Eichler explains.
Companies often don't even know what data their ‘data lakes’ contain, as it is inadequately described.
Rebecca Eichler
“Carrying out data analytics throughout the product lifecycle, gives us a better understanding of and enables us to optimize such things as products, entire factories and individual machines.”
-- Dr. Peter Reimann
Construction kit for the industrial sector
Very specific proposals for the manufacturing industry were developed by the flagship “Industrial Communications for Factories (IC4F)” project in which 14 partners collaborated having received 13 million euro in funding from the German Federal Ministry for Economic Affairs and Energy (BMWi). “That’s a lot for a computer science project,” says Dr. Pascal Hirmer, and goes on to explain why: “Industry 4.0 presents companies, especially small and medium-sized enterprises, with a huge challenge. It involves a plethora of new technologies, heterogeneous infrastructures, and compliance with data protection legislation, on top of which, the entire thing is both time-consuming and expensive. That’s why,” he continues, “we developed a reference architecture in the IC4F project, which is called iRefA (industrial Reference Architecture), and which companies can use to construct secure, robust and real-time communications solutions.” The iRefA works like a set of Lego blocks comprised of hardware and software components, network technologies, security modules and more. “Companies specify their needs in requirements workshops, and our platform suggests semi-automated building blocks that best fit the desired application. This enables the project staff to decide between the best alternatives.” The longer-term plan is to standardize the iRefA as a DIN specification.
Industry 4.0 presents companies, especially small and medium size companies, with a huge challenge.
Dr. Pascal Hirmer
Digitalization continues to make progress
As a flagship project, IC4F will spawn further research projects, which, as Mitschang explains, is important because, among other things, digitalization will continue to advance within the industrial sector. Expertise in this area is still lacking, especially in the SME sector. “That’s why we need to push the digitalization agenda and train talented people, who will go on to integrate the digital transformation into industrial applications to optimize operations.”
Institute of Parallel and Distributed Systems Application of Parallel and Distributed Systems
Universitätsstr. 38, 70569 Stuttgart
Telefon: +49 711 68588449, E-Mail