Many high-performance computing (HPC) sites support large-scale bare-metal deployments and focus00a0on delivering cost-effective solutions with great performance. In the recent past, however, HPC workloads have increasingly moved00a0to cloud environments built to support all aspects of compute and data-driven research.Join us for a discussion on building high-performance research clouds using both on-premise and off-premise resources, as well as the ability to burst workloads to the public cloud when needed.The discussion will be lead by moderators from Penn State University and Red Hat and will leverage lessons learned in several recent implementations.