Open Source Scalable Cluster File System Gains Luster

Preparing for its first commercial release later this month, the Lustre scalable cluster file system from Cluster File Systems, Inc. (CFS) has just passed several key milestones: -- Lustre powers four of the top five Linux-based supercomputers in the world, including four of the top 10 performing sites on any platform in the just-released TOP500.org list of the world's fastest supercomputers. -- Lustre went live at four key sites, including Livermore National Labs (LLNL), and the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign. -- CFS professional support for Lustre is now supporting more than 500 terabytes and over 4,000 production nodes in aggregate. Lustre did not have a single instance of data loss in any of its production sites. -- Record-breaking throughput of 11.1 gigabytes per second, the fastest sustained data bandwidth performance for any Linux cluster application. Lustre is set for general commercial release December 15, 2003. As an Open Source Linux-based storage system, Lustre is significant because its object-based architecture is expected to dramatically impact both the high-performance and the enterprise storage markets by providing significant performance and scalability increases at a fraction of the cost. CFS has embraced key open standards that are expected to accelerate Lustre adoption. Lustre is the only Linux cluster file system to be POSIX compliant and to be released under the GNU General Public License (GPL). As Open Source software, it provides strong support for industry-standard platforms and heterogeneous network environments -- building upon the success of Linux and related technologies.