Networks, multicore, and systems evolution - facing the timing beast

Embedded systems rapidly grow in several dimensions. They grow in size, from local isolated networks with simple protocols to network hierarchies to large open heterogeneous networks with complex communication behaviour. They grow in performance, from simple microcontrollers to superscalar to multicore systems with many levels of memory hierarchy. And, they expand in the time dimension by moving from static system functions to open and evolutionary functions that change over time and require new design methods and autonomous system functions. All these development contribute to an ever increasing behavioural complexity with equally complex timing. Nevertheless, the fundamental requirements to reliability and performance predictability have stayed and even been enhanced. Embedded system technology has responded with new integration methods and software architectures supported by platform control methods using new service quality metrics, and with composable formal methods that scale with system size. The talk will give an overview on this exiting scientific field and will give practical examples.