INDUSTRY

SAN FRANCISCO, CA -- Accelerating cluster adoption for supercomputing applications, Platform Computing Inc., the leader in distributed and Grid computing software solutions, released Platform Clusterware 5 at the LinuxWorld Conference & Expo in San Francisco. Platform Clusterware leverages the strengths of Platform's flagship workload management software, Platform LSF, to provide the industry's first and only cluster management solution that combines workload management and systems management into a single, integrated package. Originally announced at LinuxWorld in January 2002 for a variety of Linux distributions, this new version also runs on Sun Solaris, HP-UX and other UNIX platforms. New Clusterware partners include Dell and Microway, in addition to Atipa, Intervision, MSC Software, and RLX Technologies, which will integrate it with existing solutions to provide customized cluster management solutions to customers. "Clusterware is a unique, commercially-supported solution that integrates monitoring, scheduling and administrative functions for both Linux and Unix clusters, whereas many point solutions focus on only one aspect of clustering," said Paul Hill, vice president, marketing and business development, Platform. "As more and more applications become available in Linux environments, including bioinformatics, manufacturing and finance, this solution enables them to envision the 'cluster as server'." Clusters built on Intel-based rack-mount or blade servers running Linux represent a new approach to the expensive, proprietary supercomputers of the past. With the release of new Intel(R) Xeon(TM) and Itanium(R) processors, advances in clustering technology and the economics of standards-based technologies, use of clustering will continue to gain momentum for mainstream applications. To realize the potential of industry standard components in computing clusters, management software is needed to simplify the use and administration of systems in the cluster -- in effect treating a cluster of independent nodes as a large SMP without the shared memory. Designed for standards-based servers, Platform Clusterware enables enterprises in industries such as life sciences, research and education, manufacturing and financial services to roll out low-cost industry standards-based clusters faster, reduce operational costs, and improve computing productivity. Clusterware enables enterprises to manage the cluster as if it was one single large server, making it easy for applications to take advantage of the aggregate power of the individual nodes. By treating the 'cluster as server', enterprises can maximize on the cost-effective, high performance potential of clusters, which are estimated to achieve similar levels of performance for as little as 10 percent of the cost of proprietary SMP systems. IDC research indicates that clusters will become the dominant server architecture, with strong growth in the technical computing markets. According to Dan Kusnetzky, IDC vice president of system software research, "IDC's research indicates that organizations are increasingly looking for ways to provide more IT services while still reducing their overall IT budget. Harnessing together the power of many high performance, low-cost systems into clusters or grids appears to be a way many are considering to address both needs. Virtual processing software, such as Platform's Clusterware, appears to hit the mark for organizations needing to increase the performance of their technical applications and still be able to manage the environment as a single computing resource." "Dell's HPCC program provides customers with the power of supercomputing with all of the cost and flexibility advantages of industry-standard technologies," said Dr. Reza Rooholamini, director of clustering and operating systems, Dell Enterprise Systems Group. "Working with partners like Platform, we can provide customers with flexible and powerful solutions in data-intensive compute environments by delivering integrated cluster solutions that are tested and reliable." "Porting Platform Clusterware management software to Intel Itanium processor-based systems allows Platform Computing's customers to capitalize on cost-effective, high performance cluster solutions," said Melissa Laird, director of Intel Developer Services. "We are pleased to have Platform Computing participate in Intel's Early Access Program." Platform Clusterware includes integrated workload management, system monitoring and administration all accessed through an easy to use browser-based interface. It provides an open environment for incorporating installation, configuration or other vendor-supplied management tools. This new version supports a standards-based architecture with XML/SOAP interfaces for web services and open APIs, and runs on a variety of Linux distributions including Red Hat, Debian and SuSE Linux on Intel 32-bit or 64-bit Intel Itanium systems, and Sun Solaris, HP-UX and other UNIX platforms. Integrations are planned for Scyld Beowulf clusters and Scali cluster management technology, and the open-source Maui scheduler and will be available in Fall 2002. Platform Clusterware includes an entry-level solution for workload management with the upgrade path to meet advanced enterprise demands with Platform LSF 5. When used with both open source and commercial Grid computing stacks, such as Platform MultiCluster or Platform Globus, Clusterware enables secure, large scale sharing of resources across heterogeneous resources, departments, and locations. Platform Clusterware is available immediately through Platform's direct sales force, and its partners, including Atipa Technologies, Intervision, Microway, MSC Software and RLX Technologies. Pricing is available for both Linux and Unix systems. For more information visit www.platform.com.