HP Labs is taking part in a global Grid computing challenge
set by the organisers of the SuperComputing 2003 conference
by dedicating a 40-processor cluster to the venture.
The cluster, in HP Labs Bristol, UK, is 20 DL360 2-processor servers running Linux RedHat and Grid-enabled by Globus, which makes resources available to the Grid. The cluster is one of the largest of a total of 200 nodes running as a worldwide grid across five continents. Another HP site in Houston, USA, is contributing a nine-processor node to the challenge.
The “Global Data-Intensive Grid Collaboration”, coordinated by the University of Melbourne, Australia, will demonstrate on-demand deployment of a range of data-intensive computing applications, including natural language processing, particle physics and portfolio analysis.
The 200 Grid nodes have been made available by more than 50 academic and industrial partners for the Grid collaboration during the conference week, 17-21 November. The computing resources include PCs and workstations running Windows and Linux as well as clusters and supercomputers running Linux and other Unix-class operating systems.
Paul Vickers, Grid strategist at HP Labs Bristol, said: “We are pleased to be collaborating in this demonstration of a truly global Grid, connected using open source Grid middleware. This is the latest example of HP’s collaborative experiments in grid and utility computing.”
The 40-processor cluster is one of five in Bristol, in addition to the Utility Data Centre, with similar resources available in Palo Alto. There are plans to increase these resources, which are used both for internal IT and research.
For more information about the challenge visit: http://gridbus.cs.mu.oz.au/sc2003/
A live, interactive map of the Grid Collaboration is at: http://previn.cs.mu.oz.au:8080/gridscape-hpc1/index.jsp
and the SuperComputing 2003 home page is at http://www.sc-conference.org/sc2003/