Why it exists
Central computers run software that changes often, is deployed as packages and communicates through services. The static Classic Platform does not fit that. The Adaptive Platform, first released in 2017, standardises a dynamic, service-oriented environment.
What it is
- Runs on a POSIX operating system (e.g. QNX, Linux).
- Applications are C++ processes grouped into deployable packages.
- Functional clusters provide services: communication (ara::com), Execution Management, State Management, Update & Configuration Management (UCM), Diagnostics, Persistency, Identity & Access Management, Crypto, Log & Trace and more.
How it works
Applications describe what they provide and require; the platform starts processes according to machine states, connects services via SOME/IP, DDS or local IPC, and installs updates through UCM.
When to use it
On HPCs and powerful zone or domain computers where software is updated frequently and communicates via services. Many OEMs combine Adaptive with in-house or open-source stacks.
Best practices
- Separate platform and application release cycles.
- Version service interfaces strictly.
- Keep safety-critical control on a certified partition or safety island.
Common pitfalls
- Expecting Classic-style determinism from a general-purpose OS without timing analysis.