This project was a LLNL/UCSC collaboration. The goal was to design a scalable metadata-rich file system (MRFS) with database-like data management services. With such a file system scientist are able to perform time-critical analysis over continually evolving, very large data sets.

IMPIOUS was a LANL/UCSC collaboration to create a simulator for parallel file systems.

We've developed indexing methods to allow for the real-time indexing of an incoming data stream in order to support subsequent searches. We have also worked on scalability and reliability methods in order to allow the system to expand indefinitely with increasing performance requirements.

We present RUN, a new approach to optimal scheduling which reduces the multiprocessor problem to a series of uniprocessor problems. RUN is observed to have no more than three preemptions per job, reduces to Partitioned EDF whenever a proper partitioning is found, and significantly outperforms existing optimal algorithms.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Maecenas eleifend tellus eget hendrerit rhoncus. Mauris ornare elit eu sollicitudin bibendum. Nulla scelerisque ex faucibus, hendrerit diam id, maximus erat. Curabitur a justo dolor. Integer iaculis mauris accumsan, efficitur dui nec, porta nulla. Praesent molestie ut ex at vulputate.

Motivated by the need to provide an extensible and flexible framework beyond the abstractions provided by API libraries for files to manage and analyze large-scale data, we are developing Damasc, an enhanced file system where rich data management services for scientific computing are provided as a native part of the file system.

Solid state drives (SSDs) provide faster random I/O and use less power than hard drives, but are not yet cheap enough to substitute for all of the drives in large-scale storage systems. We present RAID4S, a cost-effective, high performance technique for integrating SSDs into RAID arrays.