Influenced by advances in data and computing, the scientific practice increasingly involves machine learning and artificial intelligence-driven methods which require specialized capabilities at the system-, science- and service-level in addition to conventional large-capacity supercomputing approaches. Expanse provides system components built around composability of data-centric applications using container coordination and integration, enabling dynamic composability of heterogeneous systems and services in scientific workflows. This talk will present the approach and the architecture of the composable systems component of Expanse. We will summarize scientific case studies that demonstrate the application of this new infrastructure and its federation with Nautilus, a Kubernetes-based GPU geo-distributed cluster. We will also describe the allocation application process for composable systems in Expanse.
