Applications › Mixed-criticality systems
Putting safety-critical and ordinary functions on one MCU saves cost, but only if the critical work stays on schedule when something else overruns. STK's commercial MCAS and MCAS4 schedulers are built for exactly that.
Functions of different criticality now share one MCU or core instead of living on separate chips.
Execution times vary. When a high-criticality task needs more time than planned, lower-criticality work must yield, and recover afterwards.
Safety standards expect you to show how non-critical code is kept from interfering with critical code.
Adaptive 2-level (LO/HI) mixed-criticality scheduler for hard real-time workloads.
4-level variant for systems with more than two criticality levels.
Automatic cascade escalation when a level overruns, and recovery when the load returns to normal.
SWRR group scheduling, with CPU share adapted at run time from per-group execution-pressure estimates.
Separate privileged and user tasks and per-task memory regions in the open-source kernel.
Combine with the commercial TrustZone interface on Cortex-M23/M33/M55/M85-class cores.
Group tasks by criticality: two levels with MCAS, up to four with MCAS4.
Give high-criticality groups their hard deadlines; let lower groups share the remaining CPU, adapted at run time.
Isolate groups with MPU regions so a faulty low-criticality task cannot overwrite critical data.
Validate behaviour on the target with SystemView traces, and use our certification support to prepare evidence for IEC 61508, IEC 62304, ISO 26262 or DO-178C.
MCAS and MCAS4 are commercial components; the other schedulers are MIT-licensed. Contact us for evaluation and licensing.
Porting, real-time tuning, certification support and commercial components are available. See services.