"No scientist can be world-class in his field and in computer science at the same time"
Heinz-Gerd Hegering worked for the Leibniz Supercomputing Centre (LRZ) for more than 50 years: In 1968 he joined as a research assistant, received his doctorate in mathematics and became a member and head of the board of directors in 1989. Under Hegering's aegis, the LRZ developed into the national supercomputing centre and moved from Munich to Garching. In addition, the professor established the computer science curriculum at the Ludwig-Maximilians-Universität. In 2008, Hegering resigned as head of the LRZ, and in 2019 he was also retired as director. 2012, for the 50th birthday of the LRZ, the scientist wrote a chronicle about the LRZ, which is now digitally available and shows a colourful kaleidoscope of stories around the data centre.
How was the chronicle of the LRZ created in 2012? Prof. Dr. Heinz-Gerd Hegering: What attracted me to the chronicle at that time was, that not only the history of the LRZ had to be written, but also the history of the development of computer science systems. When I studied mathematics and did my doctorate in this field, the subject of computer science did not even exist. Shortly after I started at the LRZ, Professor Gerhard Seegmüller became head of the board of directors. He brought back from the USA research questions about computer systems and their applications, which was new at that time. Together with a LRZ colleague I was one of the critical rehearsal listeners of his lectures, before there were no terms or definitions for IT technology and its functions. So I mutated more and more to a computer scientist. Among my first tasks at the LRZ was the development of a program library for statistical applications as well as consulting users in statistical and numerical methods. Later I was engaged in system programming, networks and communication systems, applications for personal computing, IT infrastructures and concepts for IT management.
Did you have help with the work on the chronicle? Hegering: Dr. Dietmar Täube, the former deputy director of the LRZ, and I probably started working on the chronicle for the 50th birthday in autumn 2011. He had the overview of all practical questions about the organisation, provided the information about employees, technology and structures. The LRZ was and is a unique institution in Bavaria and Germany. Compared to other university data centers, it was responsible for two universities from the beginning and thus for a much larger and more heterogeneous user group. The networking of spatially distributed users, but also the diversity of subjects of the universities required a different technical orientation and a more complex operational action. Conversely, we were often technically ahead of the game, because we always had to set up larger systems than other university computer centers. In addition, the LRZ was located at an academy, the Bavarian Academy of Sciences and Humanities (short BAdW), which at that time was mainly oriented towards the humanities, and this often resulted in much greater scope in terms of topics and procedures, which still exists today. With this special position we were able to take over inter-university tasks and also hope for other, additional sources of funding. This opened up many innovative fields of experimentation - the most recent examples are visualisation, pilot projects with energy-efficient computer architectures or the fields of quantum computing, Big Data. In any case, work on the chronicle proceeded briskly in 2011. We had a good archive at the LRZ and already collected and documented a lot about the infrastructures and systems.
What surprised you during your research? Hegering: That the LRZ-Chronicle actually created a history of IT systems in Germany, but not only from a scientific point of view, but also from the perspective of users and IT infrastructures. We have built up the LRZ chronicle system-oriented, so we rather describe the computer-technical aids of teaching and science in the course of time. This was and is interesting for computer science, because with the emergence of new systems with still unknown functionality new operational and research questions arose: How can hundreds of regional locations and independent organizations be networked? How do distributed networks work? How can a network of thousands of servers be operated efficiently and securely? Which management concepts and security measures are suitable? Which databases and software are suitable for operating thousands of PCs and clients? What is sensible to outsource to clouds? The application proximity of the LRZ, the success of the scientific projects and a professional IT management were certainly the reasons for the German Research Foundation and the German Science Council to place a national supercomputer with us around the year 2000.
You managed the LRZ from 1989 to 2008 - what was important to you? Hegering: The LRZ should gain a know-how advantage and expertise in computer-related sciences through its own research projects. This is the prerequisite for the development of innovative IT services. We wanted to pilot the latest technology and promote its use. And we wanted to be a leader in advising users: astrophysicists, natural scientists and physicians can only do good arithmetic if they are optimally supported by IT systems and IT specialists that fit their applications. Hardly any scientist can really be at the top of their field and at the same time be at the top of IT. That's why close cooperation between IT specialists and specialist scientists is so important.
Would you do or decide things differently today? Hegering: Not many. The hardware in our compute centre is already worth several hundred million euros. If the systems there are to operate well, there must be people who are not only technically competent and scientifically interested, but above all want to support our users in their projects. Services are gladly accepted, but rarely rewarded. Unfortunately, this often applies to the scientific reputation of IT application support, for example in the field of doctoral procedures, publications in scientific journals or at conferences. The problem can also occur in the LRZ, research projects may sometimes be more important than the task of keeping the systems running. So as a superior you have to take care that there is a good balance between service and scientific work. This challenge was and is certainly not trivial.
What are you doing today? Hegering: I am still closely connected to the LRZ and am in good contact with Arndt Bode, Dieter Kranzlmüller, Helmut Reiser and other members of the board of directors. Some everyday questions and problems I still get to know, but I don't think them through anymore. If I am asked for advice or if I notice something, I will say so, but of course I will not interfere. In addition, the innovation rate in our field is so high that I can no longer follow and understand everything. Maybe a new author should continue the chronicle of the 60th birthday of the LRZ and continue the developments and successes in a language that is understandable for oldies, too. Still he would have almost two years time ... (vs)