Project publication · Computer Networks
Summary
The article proposes an active IP geolocation system that uses RTT measurements from distributed monitors and machine-learning models to estimate the position of Internet nodes.
Main contribution
The contribution combines a Europe-wide measurement campaign with KNN and MLP models to improve accuracy over constraint-based methods. The system targets bounded-latency results and sufficient precision for cybersecurity applications.
Resultados y alcance
La campaña europea de medidas RTT muestra que KNN y MLP pueden mejorar la localización respecto a métodos basados únicamente en restricciones geométricas. La precisión varía con la distribución de monitores y destinos, por lo que el trabajo acota la generalización y aporta los datos necesarios para reproducirla.
Link with BoND1
Although the use case is IP geolocation, the work contributes to BoND1 with a network-observability methodology based on measurements, machine learning and experimental validation, useful for IoT infrastructures and B6G services where traffic origin and trust matter.
Project publication · Computer Networks
Summary
The paper proposes an active IP geolocation system that uses RTT measurements from distributed monitors and machine learning models to estimate the position of Internet nodes.
Bibliometric indicators
- JCR area
- Engineering, Electrical & Electronic
- JCR year
- 2025
- Scholar query
- 26 October 2026
Main contribution
The contribution combines a European-scale measurement campaign with KNN and MLP models to improve accuracy over constraint-based methods. The system is aimed at results with limited latency and sufficient precision for cybersecurity applications.
Results and scope
The European RTT measurement campaign shows that KNN and MLP can improve localization compared to methods based solely on geometric constraints. The precision varies with the distribution of monitors and destinations, so the work limits the generalization and provides the necessary data to reproduce it.
Link with BoND1
Although the use case is IP geolocation, the work provides BoND1 with a network observability methodology based on measurements, machine learning and experimental validation, useful for IoT infrastructures and B6G services where the origin and trust of the traffic matter.