Cisco Nexus Data Center Class Platform Supports Lawrence Livermore National Laboratory Linux Hyperion Project

AUSTIN, TX., SuperComputing 2008 - November 18, 2008 - Lawrence Livermore National Labs (LLNL) has selected the Cisco Nexus data center class platform as a foundation technology for its Hyperion project, a high performance computing Linux Cluster. Hyperion helps fulfill U.S. Department of Energy and National Nuclear Security Administration (NNSA) goals to provide state-of the-art computing capabilities for national security, scientific computing and discovery, and to enhance U.S. competitiveness in high performance computing.

The Hyperion project takes advantage of industry leading technologies to create the largest testbed of its kind in the world for high performance Linux cluster computing technologies and make petaFLOP/s (quadrillion floating operations per second) computing and storage more accessible for commerce, industry, and research and development. The first phase of the project is deployed, with full completion planned for March 2009. When finished, the Hyperion cluster will have at least 1,152 nodes with 9,216 cores; with about 100 teraFLOP/s peak; over 9 TB of memory; InfiniBand^TM 4x DDR interconnect and access to over 47 GB/s of RAID disk bandwidth.

To provide the high-performance unified fabric required by the project, LLNL has deployed the Cisco Nexus 7000 and Nexus 5000 Series data center switches. Deploying the new Cisco Nexus 7000 18-slot system delivers up to 15 terabits per second of switching capacity in a single chassis, supporting up to 512 10-gigabits-per-second (Gbps) Ethernet interfaces and future delivery of 40- and 100-Gbps Ethernet. The unified fabric architecture of the Nexus platform combines Ethernet and storage capabilities over a single transport, designed to provide access to all network and storage resources from anywhere in the network.

The Cisco Nexus 7000 platform fulfills the large fabric, cross-sectional bandwidth requirements of Hyperion, and supports LLNL's unique architecture that allows for easily deployed capacity expansion. The Nexus 7000 Series data plane is fully distributed and coupled with the Cisco NX-OS operating system enables zero service-disruption upgrades on production systems. This provides a seamless systems design that reduces administrative tasks and simplifies complex systems operations.

"The goal of the Hyperion Project is to provide a development, testing and scaling environment for new cluster technologies and infrastructure critical to the mission requirements of NNSA's Advanced Simulation and Computing program," said Dr. Mark Seager, LLNL project leader.

"Development of these new technologies for data center networking requires increased bandwidth, lower latency, intelligent routing, and cost effective optical cabling solutions. The emergence of data center Ethernet as a unified fabric is a big step towards addressing these issues."

Cisco Activities at SuperComputing 2008

Visit Cisco Booth # 925 for Cisco Nexus Series data center class platform demonstrations. In addition, Dr. Mark Seager, LLNL project leader, will moderate a disruptive technologies panel, "Weapons of Mass Disruption," on Thursday, Nov. 20, 2008 at 3:30-5 p.m. For more information:
http://sc08.supercomputing.org/?pg=disrupttech.html