Hybrid HPC: can we finally manage our data easily and consistently between on-prem and out in the cloud?

Public cloud use in a hybrid HPC environment remains a challenge: compute is easy, but data access comes with complications. Storage vendors have implemented various approaches to the problem of bursting data to the cloud. Some use global namespaces to stretch the filesystem others use snapshots and others shared repositories. In this talk we explore these different approaches: their strengths, weaknesses and scenarios where they best apply.