What runs here
This zone controller uses OS-managed software partitions without a hypervisor or guest operating systems. Here, MPU means memory protection unit: hardware that checks memory accesses against configured permissions.
In AUTOSAR Classic, an OS-Application groups OS objects such as tasks and interrupt service routines. It is a protection boundary, distinct from an application software component; several components may share one OS-Application.
| Component | Role |
|---|---|
| OS-Applications | Group software and OS objects into configured protection domains. |
| MPU configuration | Restricts access to code, data, stacks and supported peripheral address regions. |
| OS timing and service protection | Constrains configured execution behaviour and checks permitted OS service use. |
| Inter-OS-Application Communicator (IOC) | Provides configured communication across OS-Application boundaries, including between cores. |
| ProtectionHook | Handles detected protection violations through configured fault responses. |
These mechanisms depend on the selected OS scalability class, implementation and hardware capabilities. Memory protection does not by itself provide timing isolation. AUTOSAR Specification of Operating System
Why it looks like this
A zone controller combines gateway traffic, local sensor and actuator handling, diagnostics and power-control functions. Their criticality and workloads differ: a diagnostic request must not prevent a safety-related output from meeting its deadline.
An MCU-oriented design can retain a statically configured execution model and direct peripheral control while avoiding the integration burden of multiple guest operating systems. Its suitability depends on available memory, protection regions, timers and peripheral access controls—not simply the number of cores.
Partitioning supports an argument for freedom from interference, but does not establish it automatically. Engineers must examine memory corruption, execution delays and unintended communication, together with shared-resource failures. ISO 26262-9 addresses coexistence and dependent-failure analyses; the required evidence follows the actual safety architecture. ISO 26262-9:2018
Core separation helps allocate workloads, but shared memory buses, flash, peripherals and locks can still introduce interference. Likewise, assigning an ASIL to a function does not make its surrounding partition independently safe.
Update frequency also matters. Changing gateway or diagnostic software can alter interrupt load, memory layout or shared-service execution time. A partition boundary can help contain change, but does not make updates independently deployable or remove integration testing.
How it connects
Above this layer, application runnables execute in tasks through the RTE. Their mapping to OS-Applications determines which components share permissions and fault consequences. Communication across boundaries should use explicitly configured interfaces, with documented ownership, validity and timeout behaviour.
Below it, the OS applies protection settings using MCU hardware. Linker sections, startup configuration and driver placement must agree with those settings. Review DMA and other bus masters separately: CPU memory permissions must not be assumed to restrict their accesses.
Privileged drivers and trusted functions deserve particular scrutiny because their permissions can weaken containment. AUTOSAR OS timing protection also does not replace end-to-end deadline supervision. AUTOSAR Specification of Operating System
Other nodes connect through the communication stack and physical networks. Receiving a message into a protected partition does not establish its freshness, authenticity or safety validity. Validate external inputs before they influence local I/O or power switching.
Compared with other node classes
- HPC application processor: The grid uses a Type-1 hypervisor to separate guest environments; this MCU cell separates workloads within an AUTOSAR OS configuration.
- HPC safety island: Physically separate cores provide a different boundary, although shared resources and dependencies still need analysis.
- MCU-based ECU: Uses the same basic MPU approach; a zone controller’s gateway and I/O consolidation typically creates more shared-resource relationships to review.
- Smart node / LIN slave: This grid has no partitioning layer there; its smaller software scope is handled without these protection domains.
Best practices
- Define partitions from safety requirements, resource ownership and fault-containment needs; document every permitted crossing.
- Minimise privileged code and review trusted entry points for argument validation and bounded execution.
- Analyse worst-case interference from interrupts, locks, DMA and simultaneous activity on other cores.
- Verify generated OS configuration, linker placement and hardware permissions together on the target MCU.
- Define violation responses around controlled outputs and safe states; test containment and recovery using fault injection.
- Reassess timing and protection assumptions whenever software, routing or driver configuration changes.
Common pitfalls
- Treating separate tasks, components or cores as proof of isolation.
- Assuming an MPU also controls DMA, guarantees deadlines or eliminates shared-hardware failures.
- Granting broad shared-memory or peripheral permissions until the intended boundary becomes ineffective.
- Restarting a partition without coordinating outstanding requests, shared data and actuator state.