Résumé
Since the European Space Agency Copernicus’s launch of the Sentinel satellites in 2015, there has been a rising interest in adapting free and high-quality geospatial data to inform var- ious areas of scientific research; among them are health and biodiversity.In 2017, the UMR Espace-Dev research team at SEAS-OI Station developed an automatic data processing toolchain in Python for raw Sentinel-2 images, . Dubbed ”Sen2Chain”, this powerful tool provides researchers with a way to quickly download and host Sentinel-2 imaging data from a variety of providers, and it also enables advanced analytics on this data, including detection of changes in land cover, calculation of predefined radiometric indices (such as NDVI or NDWI), and extraction of time series data for these indices from specific areas and locations.While Sen2Chain and its web server implementation, Sen2Extract, have greatly simplified the acquisition of Sentinel 2 data, there are limitations in storage, distance, and reliability of net- work connectivity which hinder the accessibility of the data for the region of Southeast Asia. Therefore, at the KHEOBS Laboratory, we decided to host Sen2Extract in Cambodia using on premise infrastructure.In order to run the toolchain on our servers, we employed a collection of the latest technologies, such as Docker containerization, Python micro-services, and AWS Simple Storage Service pro- tocol (S3); as a result, we greatly simplified the installation process of Sen2Extract. With the introduction of Gitlab’s continuous integration technology, we further improved the resiliency of the installation. In the end, not only did we successfully adapt Sen2Extract for usage in Cambodia (https://sen2extract.kheobs.org/), but we also opened an avenue for future scaling of data related to Sentinel-2 and other large data sources.