Researchers from the Universities of Sydney, Melbourne, Adelaide and the Australian National University are demonstrating a prototype data grid -- known as the Belle Data Grid -- at a Pacific Rim Applications and Grid Middleware Assembly (PRAGMA) workshop today in a bid to demonstrate that data grid infrastructure and open source software tools can be used to automate much of the generation, storage and analysis of Belle data that is currently being done manually.
"Belle is a large high-energy physics experiment undertaken at the KEK particle accelerator in Tsukuba, Japan. It involves 400 scientists from 50 institutions around the world, mostly in the Asia-Pacific region," said Dr Lyle Winton, a researcher in the University of Melbourne's Experimental Particle Physics research groups.
"Belle is similar to the European Organization for Nuclear Research’s (CERN) Large Hadron Collider (LHC), except that it is smaller in scale and is already running -- and has produced Terabytes of data," said Winton.
Modern science experiments, such as high-energy particle acceleration experiments or proteomics experiments, can often generate Terabytes of data that needs to be accessed by hundreds of researchers around the world.
IBM has provided the team with four industry-standard Intel processor-based eServers, and is anticipating the much-hyped day when the use of grid computing in an on-demand pay-as-you-go basis pushes computing power into the realm of utility, along with water, gas, telecommunications and electricity.
IBM's director of systems sales Andrew McDougall said we'd see an evolution of grid computing along the path to a utility, moving from the scientific field to a commersialisation process that will see grid computing optimising the use of applications and infrastructure. He declined to put a time frame on being able to purchase computing power in a similar way to electricity, but said a lot of IBM's customers in the US and the UK were preparing themselves for that environment.
McDougall described todays demonstration as "a local display what research communities in Australia have been able to do and what Australian leadership can do in the field".
Winton said that after the demonstration, he expected a gradual roll-out of the system to many of the institutions involved in the Belle collaboration to begin, with a large-scale data grid with many sites throughout the Asia-Pacific region running by the end of the year.
The researchers say the Belle data grid will also serve as a prototype of a general data grid infrastructure (software and hardware) for Australia.
"It is expected that lessons learned in developing this prototype application will guide the design and development of the national computational grid that will be implemented by the second phase of the Australian Partnership for Advanced Computing (APAC) starting in 2004," said Winton.
