CNES projects library
CNES has been in the business of compute-intensive science data processing for several decades now. Initially focused on complex physical modelling (for example to devise complex mechanical parts), the agency’s Data Processing Centre today devotes most of its time to analysing and transforming data.
CNES supports projects being pursued by users or in house, or with partners outside the agency (for example, hosting their data processing and scientific research). This may involve operational tasks for a ground segment processing a space project’s payload data, but most of the agency’s activities concern upstream research and development or downstream data exploitation.
The Data Processing Centre comprises a high-performance computing (HPC) platform with more than 12,000 processing cores and 8 Petabytes of high-performance storage space shared between more than a thousand users and operational projects like CFOSat, THEIA and PEPS.
Working alongside this platform are two science data storage departments:
- The STAF data transfer and archiving department (Service de Transfert et d'Archivage de Données), CNES’s ‘memory’ tasked with preserving data and algorithms from past, present and future missions.
- Storage capacity used in particular by the PEPS project to store, disseminate and exploit data from the Sentinel satellites, which currently stands at nearly 13 Petabytes.
Users have access to a suite of software pre-installed and configured for the computing platform, as well as to software development services for configuration management and continuous integration of their code. They can also ask for personalized support from a team charged with facilitating use of the Data Processing Centre’s services. This may extend from help with accessing resources right through to code assessment and optimization, as well as promoting best practice in distributed processing and access to data or support for using software and services.
The Data Processing Centre can be accessed from outside CNES, either through a command line SSH session or via a dedicated interface to support national and international partnerships.
L-SIS-poleHPC at cnes.fr