Platform ships LSF 5 workload management solution for enterprise Grid computing

Built on a foundation of more than 10 years of distributed computing experience, LSF 5 has an open, modular architecture that delivers significant improvements in three key areas:

1. Scalability: LSF 5 boosts scalability to over 100 clusters, with over 500,000 jobs per cluster, and more than 200,000 CPUs across the enterprise. With MultiCluster 5, these capabilities can be extended across multiple sites and geographical locations through a new, groundbreaking Resource Leasing Model, which provides intelligent, dynamic, transparent, secure global sharing of resources while maintaining local and departmental ownership and control.
2. Performance: LSF 5 increases enterprise IT utilisation to 100 percent, through enhanced cross-queue fairshare policies, advance reservation, resource allocation limits, and memory reservation, which enable enterprises to utilise all heterogeneous IT resources.
3. Flexibility/Manageability: The open, plug-in architecture of LSF 5 offers a Software Development Kit (SDK) for tuning Grid policies to map to business rules, and simplifies integration with third party applications and heterogeneous technology platforms, ensuring the highest levels of standards-based connectivity for creating Enterprise Grid computing solutions. A simple, Web GUI and SOAP-based XML application interface makes global access to distributed computing resources convenient for users and administrators. MultiCluster 5 eases global sharing by providing a single system image across all sites, making management of remote resources as easy as local resources.

Over the past several months, Platform conducted extensive consultations with LSF customers, including 40 LSF 5 beta customer sites around the world. LSF 5 has been adopted by some of the world's most innovative enterprises in a variety of industries to manage their heterogeneous IT environments and build Enterprise Grids.

"Our installation of LSF 5 went extremely smoothly, 90 minutes total for both clusters, including pre-install configuration", stated Mike Bauer, Head of IT, SHARCNET, University of Western Ontario. "We are using LSF 5 with MultiCluster in production on two Red Hat 7.2 clusters, harnessing 48 and 148 CPUs that we use for both the Resource Leasing Model and Job Forwarding Model simultaneously. The new open architecture in LSF 5 is great, because it allows us to accommodate custom application integration research projects, while maintaining an extremely stable environment for running regular jobs. We look forward to extending MultiCluster across a high performance Grid in Southwestern Ontario connecting clusters at our SHARCNET partner institutions of McMaster University and the University of Guelph."

"We have been running the LSF 5 beta test version on a single cluster of over 2000 CPUs in an AFS Solaris/Linux environment", said Edwin Russell, Computer Systems Specialist, Stanford Linear Accelerator Center (SLAC). "The redesigned scheduler has worked very well and schedules jobs more efficiently under heavy load than the previous version. The transition from host-partition fairshare scheduling to the new cross-queue fairshare capability was seamless, and was mostly transparent to users."

Expanding its global reseller network to over 40 Platform partners, Platform has signed 13 new Value-Added Resellers (VARs) to deliver integrated LSF solutions for financial services, life sciences, government, academic and research and manufacturing customers, including Cnetics, Eakins Associates Inc., Enterprise Computing Solutions, EsEDA, Intervision System Technologies Inc., MSC Software, Neosummit, Nitech, Paralogic Inc., R3 Systems Inc., Stoneworks Technologi Inc., Viaken Technology and Vista Solutions Corporation. Platform LSF is also available through leading system vendors including Apple, Compaq, HP, IBM, SGI, Fujitsu, Hitachi, NEC and Sun Microsystems.

Platform LSF 5 supports both 32-bit and 64-bit versions of all major UNIX platforms, including Sun Solaris, HP-UX, IBM AIX, Compaq Tru64 UNIX, SGI IRIX, Linux for Intel and Alpha, Microsoft Windows NT, and Microsoft Windows 2000 workstations, MacOS, Cray, Fujitsu, NEC and other supercomputers, as well as AFS, DFS, and DCE environments. To provide enterprises with seamless deployment, LSF integrates with all major applications from industry leaders such as Alias Wavefront, Ansys, Fluent, Lion Bioscience, Powerllel, SAS, and Synopsys.

In addition to LSF, Platform's workload management solutions include Platform MultiCluster for Enterprise Grid computing across multiple clusters and geographic locations, Platform Parallel for managing parallel applications, Platform JobScheduler for jobflow management, Platform FTA for fast file transfer, reliability and security, Platform ActiveCluster for Windows desktop computers and Global License Broker for global license sharing. These solutions are supported by a global professional services team that provides industry domain experts, customer support, and training courses to ensure that organisations achieve maximum value from Platform's comprehensive distributed computing software solutions.