Geolocation Microservices Integration Framework for Real-Time Geolocation-Based Immigration Surveillance Support Using UML and the Sidecar Pattern

Abstract

This study proposes a geolocation microservices integration framework for immigration surveillance using UML and the sidecar pattern. The prototype was evaluated through multi-location subject testing, GNSS-based positional classification, descriptive error metrics, and comparative architectural experiments. Positional error was calculated using the Haversine formula and interpreted by subject rather than by a generalized percentage claim. The results show that six subjects were classified as Very High, one as High, and one as Moderate, with MAE of 43.34 m, RMSE of 75.20 m, median error of 16.75 m, and SD of 65.70 m. The comparative experiment indicates that microservices offer stronger modularity and fault isolation than a monolithic baseline, while the sidecar pattern improves observability and auditability without modifying the primary geolocation service. The framework is feasible as a decision-support layer for real-time immigration monitoring, subject to legal governance and further load testing.