The High Performance Storage System (HPSS) project is aimed at letting users store large amounts of data efficiently. The system is designed to be scalable, which means that there are no built-in limits to how much data HPSS can hold or how fast data can be transferred. HPSS is limited only by the storage and network technology that vendors can supply at the moment.
HPSS is an archival system, which means that it is intended to store data that won't be needed immediately, but may be needed for many years to come. To make this economically feasible, HPSS is a hierarchical file system. This means that data can be stored on a variety of media from disks to tapes, with data moving back and forth between the storage levels to make the best use of resources. For example, when a file is first stored on HPSS, it can be put on hard disks. Later, if the file is not accessed often, it can be migrated to tapes. Files are moved from disk to tape and back automatically so that the user does not need to be aware of where the data is. Moreover, the HPSS administrator does not have to take special action to see that files are moved; this happens automatically with very little help from the operator.
The HPSS project is a collaboration between IBM and the national laboratories. The system is being used by LLNL, LANL, Sandia, and Oak Ridge, as well as many other sites including universities, research laboratories, supercomputer centers, etc.