Common Infrastructure borg

Rather than building and maintaining entirely independent HPC-clusters, we aim at linking our clusters by sharing infrastructure, such as login nodes, file systems, and batch system. 

Borg is a concept of integrating separate HPC systems into a common structure. What we plan to do in the next stages:

• Using shared HOME folders (first bender, later bonna) hosted on the FDI (Forschungsdateninfrastruktur) in addition to the clusters' own HOME folders, similar to the structure at JSC
• Building a shared node that can be used to access the file systems of and submit jobs to all connected HPC clusters
• A common batch system - depending on the cluster addressed maybe pushing jobs to different (local) batch systems, like HTCluster

Access:

There is no need to apply for access to borg, being part of any application to one of the HPC clusters linked.