Multi-core hardware and software is becoming increasingly important in embedded automotive systems (EAS). While a large amount of complex algorithms, sensors, and actuators demand for more computational power, the desired integration density and real-time requirements of modern electronic control units (ECU) often necessitate truly parallel code execution. In addition, networked smart cars will require future embedded automotive software to be flexibly composed of a varying set of functions and services in order to retain the long-term interoperability of vehicles among each other. This project aims on researching the foundations and practical application of novel operating systems concepts and processor architectures for dynamically composed embedded multi-core systems with hard real-time constraints. Programming paradigms and schedulability analysis at kernel and application level is thus equally important as hardware support for connecting and isolating tasks and cores.