With the growth of the computing capacity of HPC systems, the importance of a fast access to data storage is becoming essential. The traditional solution based on NFS is not able to sustain the increasing I/O requirements. In the last years HPC centers have consequently been introducing Parallel File Systems that allow an increasing number of clients to access data in a concurrent manner. Different proprietary and open source technologies appeared in the the last years, like Lustre, Panasas, PNFS, PVFS2, GPFS and many others. Thus the skills required for system administration and the overall performance of such file systems is solution and vendor dependent. Selecting the right parallel file system and managing it right is a key element to successfully run HPC systems.
Participants will also get to visit the data center of ETH Zurich hosting Brutus, one of the largest Swiss HPC systems.