ADAS, domain controllers, and in-vehicle HPCs have pushed computing demands to the level of Linux + multi-core SoCs. The AUTOSAR Adaptive Platform provides a development framework based on POSIX, C++, and a Service-Oriented Architecture (SOA), giving high-performance automotive software a common foundation.
The two platforms target different types of ECUs, computing tiers, and development paces. Modern E/E architectures typically run both in parallel—HPCs run AP, traditional ECUs run CP, interoperating via Ethernet.
Key Understanding:AP is not the successor to CP. OEM vehicle software architectures are typically hybrid, with both coexisting—for instance, ADAS domain controllers run AP, while brake and chassis ECUs still run CP. Jotactic represents both product lines to help customers build complete solutions.
When ECU computing demands evolve from milliwatt-level control logic to requiring GPUs, multi-core CPUs, and real-time processing of cameras/LiDAR, the traditional OSEK + C + statically configured architecture is no longer sufficient.
Multi-camera / Radar / LiDAR sensor fusion requires real-time computation by GPU + multi-core CPUs.
Integrates multiple domain functions into a single high-performance SoC, replacing the distributed ECU architecture.
5G, cloud connectivity, and edge computing require large operating systems and dynamic network stacks.
Software-Defined Vehicles require a runtime environment where functions can be dynamically deployed, upgraded, and rolled back.
L3 and above autonomous driving requires massive AI / path planning computations, far exceeding MCU capabilities.
Multi-screen setups, Android/Linux application integration, and voice assistants require a general-purpose operating system.
When ECUs upgrade to HPCs, development teams no longer face MCU limitations, but rather the challenges of large-scale software engineering—complexity, real-time performance, SOA design, and ADAS safety.
HPCs run Linux, middleware, multiple applications, containers, AI inference engines... The coexistence of multiple layers, vendors, and languages introduces an integration complexity far beyond traditional ECUs.
Service-Oriented Architectures (SOA) require defining service interfaces, events, methods, versions, and discovery mechanisms. If every company defines its own interface contracts, multi-vendor integration becomes a nightmare.
General-purpose Linux cannot guarantee real-time responses, but ADAS safety functions require strict latency upper bounds. Achieving real-time performance in a POSIX environment is a non-trivial problem.
ISO 26262 ASIL-B/D requirements are difficult to meet directly on general-purpose OSs, necessitating mixed-criticality architectures, Hypervisor isolation, and certified base components.
Software-Defined Vehicles require dynamically updatable functions, demanding A/B partitioning, atomic upgrades, and failure rollbacks—mechanisms that have no equivalent in traditional ECUs.
Transitioning from C to modern C++ (17/20) introduces object-oriented programming, templates, smart pointers, package management... requiring a comprehensive upgrade of the toolchain and engineering practices.
An HPC is not an island—it must interoperate with dozens of traditional CP ECUs on the vehicle via Ethernet. SOME/IP gateways and signal/service conversions require a clear architecture.
HPC prototypes are often built using standard SDKs/tools, but production-grade maintainability, safety certification, and long-term support require industrial-grade frameworks, not laboratory-style integration.
The AUTOSAR Adaptive Platform provides a standardized runtime environment, APIs, communications, and lifecycle management, turning HPC automotive software development from an 'integration adventure' into an 'engineering practice'.
Standard C++ APIs such as ARA::COM (Communication), ARA::PER (Persistency), ARA::LOG (Logging), ARA::SM (State Management), and ARA::DIAG (Diagnostics) provide application developers with a unified interface, eliminating the need to deal with underlying OS differences.
Defines service contracts via SOME/IP and ARA::COM: interfaces, events, methods, fields, and versioning. OEMs and Tier1s share a common design language, eliminating the need to negotiate multi-vendor integration from scratch.
EB corbos Linux provides an AP-verified Linux distribution, integrating PREEMPT_RT, CPU isolation, and real-time scheduling mechanisms to achieve predictable response times on a general-purpose OS.
Isolating ASIL functions and QM applications on the same SoC via a Hypervisor, combined with ISO 26262 certified AdaptiveCore foundational components, makes ADAS safety functions feasible on HPCs.
Update & Configuration Management（UCM）標準化 OTA 部署流程——封包驗證、原子升級、失敗回滾。符合 UN R156，與 CP Bootloader 形成端到端 FOTA Solutions。
Unlike CP's static configuration, AP applications can be started, stopped, and updated at runtime. This turns the SDV visions of 'Function Subscription' and 'Vehicle as a Platform' into technically achievable goals.
Both AP and CP belong to the AUTOSAR standard. SOME/IP gateways can convert CP signals to AP services and vice versa. HPCs and traditional ECUs work collaboratively under a unified architecture.
EB corbos Studio provides SOA design, ARXML editing, model validation, and code generation. It offers complete tool support for engineering practices from PoC to mass production, eliminating manual integration.
Elektrobit EB corbos is a complete solution for the Adaptive Platform, covering the OS, middleware, Hypervisor, and development tools. Jotactic provides local technical support and integration services.
A complete implementation of the AUTOSAR Adaptive Platform middleware, providing core functions such as ARA APIs, SOME/IP communication, state management, and diagnostic services. Serves as the foundational runtime environment for HPC application development.
A Linux distribution designed specifically for automotive AP mass production, integrating real-time extensions, secure update mechanisms, and long-term maintenance commitments, providing Linux with industrial-grade reliability in in-vehicle environments.
A lightweight automotive Hypervisor that allows ASIL safety workloads and QM applications to coexist on the same SoC, providing necessary workload isolation and resource management mechanisms for HPC platforms.
An integrated development environment for the Adaptive Platform, providing SOA service modeling, ARXML editing, Manifest configuration, code generation, and consistency validation, serving as the core tool for AP application development.
Jotactic provides EB corbos product consulting, licensing, technical training, and HPC integration support, helping your team successfully move from PoC to mass production.