Software-defined vehicles, which rely on zonal architectures, run multiple applications on a small number of control units. This requires deploying several applications on a single platform, which introduces significant challenges in ensuring strict time boundaries for safety-critical operations due to shared physical resources and full-stack operating systems. To address these challenges, this paper proposes a time-aware scheduler implemented on an FPGA within an MPSoC for the widely adopted DDS standard, allowing time-critical data to bypass the slow and unpredictable OS-level networking layers. The scheduler enables prioritized processing of time-critical data within fixed and predictable time slots while preventing starvation of low-priority data. Moreover, we introduce a method to define QoS profiles, which allow forwarding time-critical traffic in guaranteed time slices to automotive applications running on the MPSoC’s CPUs. That way, we guarantee strict real-time requirements with minimal latency and an overall jitter below \(2\,\upmu s\) . Finally, we outsource DDS security operations to a hardware-implemented cryptographic algorithm and demonstrate a PUF-based key extraction method with 99.33% reliability. The approach enables highly deterministic processing of DDS traffic with integrated security by design.