In general, the main focus of research and teaching at the chair lies in the area of system software, so those programs responsible to make a computer operational and are commonly referred to as operating system. Both centralized and distributed computing systems are considered, which operate reliably under certain application-specific constraints (regarding time, energy, and/or space) and whose computing units themselves are single, multi or many-core systems. For the latter in particular, scalable solutions for managing simultaneous, multi-threaded processes play a major role. Accordingly, the subject areas of operating systems and distributed systems form the two main pillars of the chair, braced by work in the topics of real-time systems and energy-aware systems.
One important research topic is the adaptability of system software “upwards” to the user application as well as “downwards” to the hardware. Interoperability, portability and the degree of specialization (esp. extension, replacement, shrinking) are the foundation of software used to control all incurring procedures in a computing system in an application-oriented fashion. In this context, open interfaces and the specification of functional and non-functional properties are of particular importance, and both together form the chair’s perception of an “open system”. This is approached mainly by using methods and techniques for constructing the software instead of relying on standards. The goal of our work in this area is the improvement of the adaptability of functional and non-functional properties while maintaining control of the accompanying complexity.
In addition to the construction of system software, the chair also deals with the topic of static and dynamic program analysis for a) collecting data and control-flow dependencies in (non-sequentual) programs and b) generating a priori knowledge about non-functional characteristics such as upper limits regarding execution time (worst-case execution time, WCET) or energy demand (worst-case energy consumption, WCEC). The purpose of this knowledge gained by the program analysis is to minimize or hide latencies in the system, maintain timeliness and achieve an overall resource- and energy-efficient operation of the computing system.
The guiding principle of all our work is to always show and prove the sustainability of the researched concepts in terms of actual system software. A characteristic of our fundamental research at the chair is to apply a pronounced concreteness to our work. At the end, the goal is the realization of a concrete implementation or the integration in or adaption to an existing system. The before/after comparison of the developed system (or part of it) with existing solutions represents therefore the core of our evaluation of the concepts emerged from our research.