Integrating Scientific GUI Applications into Autoscaled Science Gateway Environments

We present an application of a new capability on the NSF Jetstream2 cloud that enables scalable access to established scientific desktop applications in the atmospheric sciences, including the Unidata Integrated Data Viewer (IDV), the AWIPS CAVE client, and the LROSE Hawkeye tool. By leveraging OpenStack Magnum's Kubernetes autoscaling clusters-as-a-service, in combination with JupyterHub and VNC technology, the NSF Unidata Science Gateway dynamically provisions server resources to meet user demand, allowing these desktop applications to run efficiently in a cloud environment. This approach makes it feasible to support resource-intensive GUI tools in scenarios where static provisioning is costly or infeasible. We describe our deployment strategy, emphasizing that this model offers a workflow for integrating desktop applications with notebook-based gateway features and applicable to other domains with similar needs.