FabAID is built on two open-source platforms, developed in the open for decades and trusted at national scale — the Pelican Platform for data, and the HTCondor Software Suite for computation. Together they let any researcher connect any repository to computing capacity anywhere.
Pelican federates data repositories into a single namespace and delivers their objects — through a national network of caches — right next to the compute that needs them. It is the software foundation of the Open Science Data Federation. In FabAID, we are evolving Pelican to meet the demands of AI-driven science: tighter ecosystem integration, metadata extraction and discovery, lightweight ways to connect smaller repositories, and improved staging management.
The HTCondor Software Suite manages and executes high-throughput workloads whether on a laptop or hundreds of thousands of cores. In the fabric, HTCondor is the workload engine powering FabAID's Access Points and data workflow services, scheduling the jobs that generate, transform, and move data across the federation.
Both platforms are free and open source, built by CHTC at the University of Wisconsin–Madison and used worldwide.