Quality of Service for Integrated Modular Avionics (IMA) on Multicore Processors using a Safety Net Architecture

Future avionic applications will require higher computation performance while at the same time a reduction in space, weight and power is needed. This need is shown for example in the concept of the Airbus Vahana aircrafts which will be ultra light-weight electrical helicopter-style vehicles providing novel autonomous urban transportation. In comparison to current aircrafts the avionics have to be much smaller and lighter while at the same time provide sufficient computation performance to compute not only the flight control data similar to current airplanes but additionally compute the complex algorithms for autonomous flying. One solution for these demands is the consolidation of flight applications currently running on multiple single core processors on a small number of multi core processors. Furthermore, legacy applications shall be reused without major modifications. The contribution of this paper ia a Safety Net approach for IMA multicore systems executing a critical IMA application on one core concurrently to other less critical applications on other cores. The Safety Net continuously tracks the progress of the critical IMA partitions or the complete IMA system. Furthermore, it triggers countermeasures if an acceptable delay is exceeded. Our proposal allows tracking the progress of an application without any modification of the application code. Therefore, it is well suited for migration of safety critical legacy IMA applications to one core of a multicore system.