I've been working on an even faster storage solution to better support the I/O needs of our computational activities. Today I'm pleased to announce the availability of the NOBACKUP storage service. Based on the combination of the Lustre parallel filesystem with the ZFS filesystem, NOBACKUP is specifically designed to handle multiple simultaneous data needs from in-house computational processes while leveraging transparent compression for efficient utilization of our storage infrastructure. As its name more than implies, there aren't any backups of NOBACKUP as it's intended only for data downloaded from external authoritative sources for analysis, or to store temporary & intermediate data during active computation.
You can find out more about NOBACKUP on our system page at either: