Automotive Software-Defined Vehicle (SDV) Gateway Blueprint: SOAFEE Architecture, OTA, & Edge Intelligence
Software-defined vehicles force a fundamental change in how automotive engineering systems are designed and maintained. Vehicle functionality and software continue to change through updates, recalibration, diagnostics, and data-driven behavior throughout the vehicle’s operational life. These changes occur while the vehicle remains in use, often for ten to fifteen years, under strict safety and reliability constraints.
This shift aligns with the rapid expansion of the software-defined vehicle market, which reached approximately USD 290.33 billion in 2025. The market is projected to grow from USD 390.29 billion in 2026 to beyond USD 4,162 billion by 2034, with a Compound Annual Growth Rate (CAGR) of 34.43%, driven by software-centric autonomy, safety systems, and connected vehicle platforms.
This growth is driven by the need for vehicle architecture that can sustain long-term software evolution without compromising safety, reliability, or operational control.
MosChip has recently introduced its ProductXcelerate Blueprint for an SDV Gateway, defined as a reference architecture for software-defined vehicle platforms.
This blueprint presents an SDV gateway designed as a persistent control layer rather than a passive networking element, reflecting established automotive hardware engineering practices for long-lifecycle systems. Built on an automotive-grade hardware foundation and aligned with SOAFEE and Adaptive AUTOSAR, it integrates OTA lifecycle management and edge AI-based intelligence into a single, coherent system.
Vehicle-level gateway control function
In contemporary vehicle architectures, functional behavior is distributed across multiple ECUs and software layers. Software components interact through in-vehicle communication networks and evolve through controlled software updates. Within this context, the gateway serves as a stable system reference point. It provides visibility into communication traffic, software versioning, update status, and system health, while remaining non-intrusive to time-critical vehicle functions.
This role establishes the gateway as a vehicle-level control function. The gateway does not host vehicle application logic. It provides coordination for inter-system communication, software update propagation, and lifecycle state observability across the vehicle platform.
Blueprint Scope and System Composition
The SDV Gateway Blueprint consists of three primary elements:
- NXP S32G System-on-Chip (SoC) vehicle network processor
- SOAFEE-aligned gateway software stack
- Functional layers implemented through MosChip’s DigitalSky GenAIoT and AgenticSky ControllerCore
Together, these elements define a reference gateway architecture intended to support system coordination, software lifecycle management, and data exchange within software-defined vehicle platforms.
The hardware platform defines the execution foundation for gateway operation by providing processing cores, in-vehicle networking interfaces, hardware-enforced isolation mechanisms, and on-board memory and storage. The gateway software stack defines execution boundaries, deployment structure, and OTA update handling on top of these hardware resources. GenAIoT and AgenticSky add operational functions that interpret system state and apply corrective actions within the gateway environment.
Hardware architecture for sustained operation
The SDV Gateway Blueprint accelerator is built using the NXP S32G System-on-Chip (SoC), selected for its balanced support of networking, real-time control, and security within a single automotive-qualified device. The hardware design emphasizes sustained and predictable behavior.
The gateway supports up to 8 Arm Cortex-A53 clusters alongside dual or quad-core lockstep M7 controllers. This configuration enables a clear separation between application-level workloads and time-sensitive control functions. Diagnostics, analytics, and update management operate concurrently with real-time networking without contention. 16 resource-isolation domains further ensure deterministic execution for mixed safety-criticality workloads, and can be partitioned using a hypervisor.
Networking capabilities help the gateway play its role as a convergence point. Support for up to 20 CAN and CAN-FD interfaces, FlexRay, and 3X 2.5-Gbps Ethernet interfaces, while hardware acceleration is handled by the low-latency communication engine, and the packet forwarding engine enables efficient message routing, filtering, and forwarding under sustained load, preserving timing discipline across networks.
Memory and storage choices support continuous operation. With 4 GB of LPDDR4 and 32 GB of eMMC, the gateway can stage OTA images, process telemetry, retain configuration state, and execute parallel software workloads. The design supports operation across automotive temperature ranges, allowing the gateway to maintain its state and recover through power cycles and network interruptions.
SDV gateway stack from SoC hardware to AI intelligence layer
SOAFEE-aligned software structure
The gateway is implemented as a SOAFEE-compliant gateway that manages data exchange between vehicle domains, ECUs, and cloud systems.
The gateway software stack follows SOAFEE principles to provide a consistent structure across development, deployment, and operation. The software design addresses execution boundaries, modular deployment, and lifecycle governance.
The software environment separates timing-sensitive functions from general-purpose services, including workloads associated with edge AI-based behaviour and anomaly models. Safety-relevant workloads execute on deterministic paths, while Linux-based services operate in controlled environments. Resource access is explicitly defined.
With the SOAFEE, the software components are deployed as discrete, versioned units with defined dependencies and validation requirements. OTA updates are applied incrementally rather than as monolithic replacements. Each update includes compatibility information, rollback conditions, and health checks.
The software structure supports reuse of the gateway framework across vehicle programs and hardware revisions. Platform variation is constrained within defined boundaries.
Integration of MosChip DigitalSky GenAIoT and AgenticSky
MosChip GenAIoT manages four functional areas: IoT and Connectivity, Cognitive, Unified Automation, and Digital Dashboards.
The IoT and Connectivity functions include CAN data ingestion and COVESA Vehicle Signal Specification (VSS) mapping. The Cognitive functions include behavior and anomaly models executed as edge AI workloads. Unified Automation includes OTA orchestration and regression automation. Digital Dashboards present fleet lifecycle information and proactive insights.
While the AgenticSky introduces the ControllerCore, which applies self-healing logic, diagnostic modes, and optimization guidance to ensure safe, adaptive gateway operation.
Supported gateway-level use cases
- OTA lifecycle management and controlled recovery: In a fleet-wide software update scenario, the SDV gateway coordinates staged OTA update execution at the vehicle level, validates update readiness, and monitors execution state during deployment. GenAIoT manages OTA orchestration and regression automation, including update sequencing and validation workflows executed at the gateway, with update packages and rollout policies provisioned from cloud services such as AWS, and execution status reported back to the cloud. When update execution deviates from expected behavior, AgenticSky’s ControllerCore applies diagnostic modes and controlled recovery actions, such as rollback or entry into diagnostic states, within defined operational boundaries.
- ECU diagnostics and health monitoring: During normal vehicle operation, the gateway supports continuous observation of ECU behavior and in-vehicle communication patterns through CAN data ingestion and COVESA VSS mapping. Cognitive functions analyze behavior and anomaly trends over time to identify deviations from expected operation, while AgenticSky provides diagnostic modes that support fault isolation and guided troubleshooting without participating in software update execution.
Deployment scope
The SDV Gateway Blueprint supports multiple deployment scenarios, including vehicle platforms, fleet systems, and SDV validation environments. Vehicle platforms apply the gateway to continuous software update management, while fleet operators use it for lifecycle monitoring and predictive maintenance. Automotive Tier-1 suppliers and OEM laboratories use the blueprint as a reference platform for SDV validation and experimentation. Deployment relies on the deterministic hardware foundation, SOAFEE-aligned software structure, and integrated OTA lifecycle management.
Implementation and customization of the blueprint are supported through MosChip’s ProductXcelerate framework. Within this framework, MosChip applies hardware engineering capabilities across automotive SoC integration, gateway board design, in-vehicle networking, and system-level validation.
MosChip, with over 25 years of experience, provides Product Engineering Services, including development and integration across the product lifecycle, including hardware, embedded software, connectivity, digital platforms, and edge AI. The scope includes board design, high-speed interface implementation, power and clock architecture, and board bring-up. Platform enablement covers BSP development, device driver integration, middleware configuration, and operating system adaptation, while connectivity support includes CAN, LIN, Ethernet, diagnostics integration, and cloud interface integration.
