The usability of any HPC system with distributed memory depends crucially on various software layers, that create the supercomputer consciousness of the huge collection of hardware components to act coherently on the given task.

Our group works on enhancement of certain components of this stack especially in the fields of resource management and scheduling. This work is done in close cooperation with partners from academia and industry and with special focus on the necessary extensions to adapt today's middleware solutions to the challenges evoked by the Modulare Supercomputing Architektur (MSA).