How Cesium is Using glTF to Build 3D Map Experiences: New 3D Tiles Now Supports Streaming of Massive Geospatial Datasets

As the “JPEG of 3D,” glTF™ from Khronos® has made a big impact in the world of 3D, enabling the efficient transmission and loading of 3D scenes and models by applications. Cesium, a platform for creating 3D applications that are fast, flexible, and based on real-world geospatial data, has used glTF extensively to further its mission of empowering developers and data providers to build web-based 3D map experiences, and now Cesium has teamed with Uber to continue expanding its 3D Tiles ecosystem, built on glTF.


Cesium introduced 3D Tiles in 2015 as an open specification for streaming massive heterogeneous 3D geospatial datasets. Expanding on Cesium’s terrain and imagery streaming, 3D Tiles has been used to stream 3D content (including buildings, trees, point clouds, and vector data), with the ultimate goal of improving streaming and rendering performance of massive heterogeneous datasets.

Since its inception, Cesium’s 3D Tiles has used Khronos’ glTF, the royalty-free specification for the efficient transmission and loading of 3D scenes and models by applications. When new tiles come into view or a greater level of detail is needed, 3D Tiles streams new glTF models. By minimizing the size of 3D assets and the runtime processing needed to unpack and use those assets, glTF defines an extensible, common publishing format for 3D content tools and services that streamlines authoring workflows and enables interoperable use of content across the industry.

Besides its 3D Tiles ecosystem, Cesium has long been an active contributor to the Khronos community. In fact, Cesium CEO, Patrick Cozzi, also serves as Chair of the Khronos 3D Formats Working Group that manages the glTF standard. A longtime champion of open-source technology, he is co-creator of glTF, creator of the Cesium WebGL™ engine, and author, editor, and contributor to ten books on mapping, computer graphics, and games, including WebGL Insights, OpenGL Insights, and 3D Engine Design for Virtual Globes. With Cozzi, the team at Cesium that designed 3D Tiles has used WebGL, the Khronos web standard for interactive 3D graphics in your browser, since its very beginning—and before that, they used Khronos’ OpenGL®, the most widely adopted 2D and 3D graphics API. By using glTF in their 3D Tiles ecosystem, Cesium has been able to streamline the process of receiving a 3D Tile and rendering it with WebGL, making the pipeline more accelerated and simplified and, in turn, minimizing client-side processing.

Cesium cites Khronos’ glTF as a normative reference for its runtime efficiency, interoperability, and strong software ecosystem, and the platform’s use of glTF has proven key to bringing the computer graphics and geospatial communities together. Earlier this year, a liaison was formalized between Khronos and the Open Geospatial Consortium (OGC), and 3D Tiles was adopted as an OGC Community Standard. With glTF’s ability to support modern graphics techniques and 3D Tiles’ ability to stream and render massive 3D content, Cesium is using glTF to help set a new bar for geospatial visualization.

In continuing their mission to build web-based 3D map experiences, Cesium has recently announced that 3D Tiles can now be loaded using, Cesium’s open source suite of loaders for big data visualization. The 3D Tiles loader in selects which tiles in a 3D tileset should be loaded for a given view; then,’s glTF loader decodes those tiles. This means that 3D Tiles is now supported by visualization and analysis engines (such as those used by Uber), enabling those engines to stream massive 3D geospatial datasets.

3D Tiles is critical for the geospatial community. With increasing numbers of autonomous cars, drones, and other sources generating increasing amounts of 3D geospatial data, the community needs a platform that can efficiently stream these ever-growing heterogeneous 3D geospatial datasets to build 3D map experiences. With support for glTF and 3D Tiles in an increasing number of visualization and analysis engines, 3D geospatial data is becoming more accessible and usable across multiple platforms than ever before.

Authored by