Assessment of Display Performance and Comparative Evaluation of Web Map Libraries for Extensive 3D Geospatial Data
MapLibre GL JS processes large point clouds 7,000x faster than CesiumJS in new performance study.
A team of Japanese researchers led by Toshikazu Seto has published a comprehensive performance evaluation comparing two leading web mapping libraries for handling extensive 3D geospatial data. The study, focused on infrastructure critical for Japan's digital society, tested CesiumJS and MapLibre GL JS using real-world datasets from the VIRTUAL SHIZUOKA and PLATEAU building model projects. Employing standardized formats like 3D Tiles 1.1 and Mapbox Vector Tiles (MVT), the research provides crucial benchmarks during the transition to next-generation standards like WebGPU.
The results reveal a stark performance divide based on data type and scale. For visualizing 3D building models using MVT, MapLibre GL JS delivered optimal performance with a First Contentful Paint of 0.8 seconds and zero Total Blocking Time. Most dramatically, for processing large-scale point cloud data, MapLibre completed tasks with a TBT of just 3 milliseconds, while CesiumJS struggled with a TBT of 21,357 milliseconds—a difference of over 7,000 times. The study establishes a reproducible framework for performance evaluation, offering clear, data-driven selection guidelines for developers and urban planners choosing technologies for digital twin applications and complex 3D web visualization projects.
- MapLibre GL JS processed large 3D point clouds with a Total Blocking Time of 3ms vs. CesiumJS's 21,357ms—over 7,000x faster.
- For MVT-based 3D building visualization, MapLibre achieved near-instant loading with a First Contentful Paint of 0.8 seconds and 0ms Total Blocking Time.
- The study provides a reproducible performance evaluation framework using Core Web Vitals, guiding technology selection for digital twin and urban planning projects.
Why It Matters
Provides concrete performance data for developers building complex 3D web applications, digital twins, and smart city infrastructure.