The GeoVER project set out to create a flexible, geographically aware warning system that could deliver real‑time traffic information to a wide range of users. Central to the design was the integration of the DisGB distribution platform, developed by the Mobile Cloud Computing research group, with a newly introduced serverless component called tinyFaaS. tinyFaaS, also created by the same research group, was published in the EuroSys 23 conference and provides a lightweight, open‑source framework for executing GeoVER’s core functions without the need for dedicated servers. By embedding tinyFaaS into the GeoVER architecture, the system can generate and route geowarnings on demand, scaling automatically with the number of incoming data streams.

GeoVER’s architecture builds on the classic publish/subscribe model but extends it with geofencing capabilities through GeoBroker. GeoBroker allows publishers to tag messages with a geographic boundary, ensuring that only subscribers located within that boundary receive the notification. DisGB then uses these geospatial tags to route messages efficiently across the network, reducing latency and bandwidth consumption. The GeoVER‑Bridge component connects DisGB, tinyFaaS, and end‑user interfaces, providing a unified API for both data ingestion and delivery. This modular design means that the same core can be reused in non‑traffic domains, such as environmental monitoring or emergency response.

A key scientific contribution of the project was the development of a warning generator that processes raw sensor data—such as traffic counts, weather station outputs, and flight‑plan feeds—to produce actionable geowarnings. The generator applies filtering and prioritisation rules that consider both the topic of interest and the user’s location, ensuring that only relevant alerts are transmitted. Experimental evaluations, conducted in collaboration with the project partners, demonstrated that the system can deliver warnings with sub‑second latency under realistic load conditions, a performance metric that meets the stringent requirements of real‑time traffic management.

In addition to the backend, the project delivered a prototype for immersive visualization using extended reality (XR). Developed by Deutsche Telekom, the XR prototype allows users to view geowarnings in a spatial context, overlaying alerts onto a 3‑D map of the surrounding area. This feature was tested with pilots at the Schönefeld airport, where pilots received location‑specific information about nearby weather and traffic conditions. The XR interface also supports multi‑user collaboration, enabling teams to discuss and respond to alerts in a shared virtual space.

The project’s technical achievements were complemented by a comprehensive set of deliverables. All source code, including the tinyFaaS implementation, the GeoBroker extensions, and the XR client, is available on GitHub under open‑source licenses. A series of scientific publications, presented at conferences such as EuroSys and in peer‑reviewed journals, detail the system’s architecture, performance evaluation, and potential extensions. The project also produced a final report and a set of best‑practice guidelines for deploying geographically aware warning systems in other sectors.

GeoVER was a joint effort between the Technical University of Berlin and Deutsche Telekom. The university led the development of the warning generator, the integration of DisGB, and the overall system architecture, while Deutsche Telekom focused on the XR visualization and user‑interface components. The project ran over a multi‑year period, with two major milestones: the first milestone delivered the warning generator and XR prototype, and the second milestone completed system integration, sensor data ingestion, and a full deployment. Funding was provided through a European Union grant, enabling the collaboration to bring together academic research and industry expertise to produce a robust, scalable warning platform.