Mobility Trajectory Data Stream Processing Beyond the Cloud
The increase in Internet-of-Things (IoT) sensors mounted on moving objects has resulted in continuous spatiotemporal data streams. Nevertheless, current mobility stream processing systems remain inadequate, as they are only optimized for the cloud and cannot effectively cope with IoT workloads. General-purpose stream systems that operate across edge–to–cloud settings do not support spatiotemporal operations. In contrast, existing spa tiotemporal libraries are primarily designed for historical analysis rather than real-time streaming workloads. We propose Mobility Nebula, a system that supports spatiotemporal stream processing across the edge–to–cloud continuum. We evaluate the system in a real-world deployment with the Belgian railway operator (SNCB), where train-mounted edge devices ingest sensor data and execute real-time mobility queries such as brake system mon itoring and high-risk zone proximity monitoring directly at the edge. In our evaluation, MobilityNebula sustained near real-time latencies for queries while ingesting a 20k events/s stream on an edge device such as a Raspberry Pi 5.
