Innovative Grid Applications Highlighted at
Fourth PRAGMA Workshop in Melbourne, Australia
June 16, 2003
Media Contact: Greg Lund, SDSC Communications, 858-534-5143, <email@example.com>
Technical Contact: Teri Simas, PRAGMA Program Manager, 858-534-5034, <firstname.lastname@example.org>
Melbourne, Australia -- Researchers from across the Pacific Rim met June 5-6 for the Fourth PRAGMA Workshop, the semi-annual workshop for members of the Pacific Rim Application and Grid Middleware Assembly. More than 70 application and grid experts from 13 Pacific Rim member organizations continued their efforts to build sustained collaborations and to demonstrate the advancing technology of grid computing via scientific applications.
Hosted by Monash University and the Australian Partnership for Advanced Computing (APAC), the workshop was chaired by Professor David Abramson of Monash University in Melbourne, and co-chaired by Dr. Fang-Pang Lin of the National Center for High Performance Computing (NCHC) in Hsinchu, Taiwan.
PRAGMA showcased a comprehensive agenda of distinguished speakers, real-time demonstrations, and working group activities.
"In this meeting we demonstrated different application-middleware technology uses," said Abramson. These included:
* Nimrod/G, a tool for distributed parametric modeling, and GAMESS, a quantum chemistry code, conducted by Abramson and by Professor Kim Baldridge and Dr. Wibke Sudholt of SDSC, the University of California, San Diego (UCSD), and the National Biomedical Computation Resource (NBCR);
* The Belle Data Grid Project, which demonstrated middleware packages such as Nimrod and Gridbus. Gridbus is an open source cluster and grid middleware technology for the management of distributed computation, data and applications, and was applied to a high-energy physics application conducted by Lyle Winton of the University of Melbourne;
* Gfarm, a software package of the Data Grid project (a collaboration among High Energy Accelerator Research Organization (KEK), National Institute of Advanced Industrial Science and Technology (AIST), the University of Tokyo, and Tokyo Institute of Technology) that is designed for constructing petascale to exascale (trillion gigabyte) parallel filesystems; the demo application of astronomical data analysis was conducted by Osamu Tatebe of the AIST;
* Ninf-G, a grid technology that allows users to access various resources such as hardware, software, and scientific data on the Grid. A climate simulation using Ninf-G was conducted by Yoshio Tanaka (AIST); it utilized 185 processors contributed by PRAGMA and Asia-Pacific Grid (ApGrid) member organizations including AIST, the Korea Institute for Science and Technology Information (KISTI), Kasetsart University, NCHC, the Tokyo Institute of Technology (Titech), Doshisha University, and the University of Hong Kong.
A tutorial was conducted using the locally developed Nimrod/G system, a project supported by the Distributed Systems Technology Center (DSTC), GrangeNet, and Monash University. Nimrod/G makes it very easy to create large parameter sweep and search applications, and has been applied to a range of scientific and engineering applications over the years. Nimrod scavenges compute cycles from a wide variety of platforms, ranging from simple workstations to high performance clusters and vector supercomputers. It also incorporates quality-of-service metrics that make it possible to set deadlines and budgets for computation.
In the tutorial, Abramson and his colleagues demonstrated Nimrod's ease of use, and launched an application of Nimrod to Quantum Chemistry using the GAMESS package. The aim was to use GAMESS for parameterizing a quantum chemical pseudo-potential, which can ultimately be applied to model large molecular systems such as proteins. This demonstration involved resources at Monash University, Victoria Partnership for Advanced Computing, Kasetsart University, AIST, Cray Inc Japan, and the University of California, San Diego and the San Diego Supercomputer Center. After running overnight, the results were interpreted by Sudholt, a postdoc of Baldridge and J. Andrew McCammon.
"These results significantly reinforced our early work in computing pseudo-potentials for organic functional groups", said Dr. Sudholt. "Most importantly, they allow us to explore many more parameter combinations than we had been able to do in the past. Future plans involve incorporation of such infrastructure for general study of complex reaction processes where one has involved setups, considerable numbers of similar runs, and investigations where one wishes to investigate effects of several parameter investigations. Such applications might range from understanding complex biochemical reactions, to design of new drugs, to basic structure-function relationships in materials, since the associated software and infrastructure can facilitate both greater control over simulations of large molecular systems, as well as investigations of large functional group spaces for molecular systems.
"Taken collectively, these demonstrations show the power of the grid, but they also show us how difficult the grid is to use. It is precisely for this reason that PRAGMA was started, to build teams of researchers and an expertise base to make the grid fulfill its promise," said Philip Papadopoulos, co-PI of the NSF award that funds PRAGMA.
Professor John O'Callaghan, Executive Director of APAC and a member of PRAGMA's Steering Committee, indicated that the workshop was a wonderful opportunity for exposing international researchers to grid projects in Australia. He said that "the workshop has strengthened collaboration between Australian researchers and other PRAGMA members." The applications described included:
* The FilmEd project, which concerns broadband access to a wealth of high quality and unique film and video content within unique Australian moving image archives;
* The Australian Nimrod Testbed project (ANT), which is rolling out the Nimrod/G software for use by Australian scientists and engineers;
* The Belle Data Grid project, which involves accessing data from the KEK particle accelerator in Japan for high-energy physics experiments;
* The Australian Virtual Observatory project to archive, retrieve and access data from a variety of Australian optical and radio telescopes;
* The NANO (Nanostructural Analysis Network Organization) Virtual Instruments project, which is starting to provide remote access to expensive microscopes at the Universities of Queensland, Melbourne, New South Wales, and Western Australia;
* The ACESS Major National Research Facility Geo Sciences project, which is creating a distributed simulator for modeling earth processes; and
* The Atmospheric Sciences grid workflow project, which is demonstrating the application of Grid technologies to atmospheric and pollution modeling.
All of these applications are supported by GrangeNet, a multi-gigabit Australian network linked to overseas research and education networks (www.grangenet.net).
Grid technology has also been demonstrated to be useful in battling real-life situations. Dr. Fang-Pang Lin and his team at NCHC have most recently been involved on the ongoing efforts to apply Access Grid Technology for medical personnel quarantined due to SARS in Taiwan. This international effort, coordinated by PRAGMA, has helped to solidify international collaborations within the Access Grid community. One application discussed at this workshop involves the Access Grid and Data technologies used to assist NCHC in its efforts to quickly deploy a network to connect hospitals in Taiwan with experts outside the quarantined areas. (For information on NCHC's SARS Combat Task Force, please visit http://antisars.nchc.gov.tw/.)
It was reported at the PRAGMA meeting that the epidemic of SARS in Taiwan seems to be improving, with no new cases reported for several days. "We are delighted by the support we have gotten via PRAGMA, and pray that the end is in sight," Dr. Lin said. "We at NCHC are looking forward to hosting the next meeting on October 22-23, 2003. At that time we hope to examine how our recent experience can be expanded to other institutions, and how to be better prepared for other emergency responses."
On the second day of the meeting, the working groups reviewed plans for the coming year. The groups include Telescience (led by Shinji Shimojo of the Cyber Media Center at Osaka University and Fang-Pang Lin of NCHC), Biosciences (led by Larry Ang of the BioInformatics Institute (BII), Singapore), Data Grid (led by Osamu Tatebe of AIST), and Resources (led by Kishore Sarkahar of BII and Mason Katz of UCSD). In addition, a new Middleware working group was organized by Dr. Andrew Wendelborn of Adelaide University, to focus on the software that sets up connections between applications and resources. The groups plan to demonstrate their capabilities at the SC2003 high-performance computing and networking conference, to be held in Phoenix, Arizona this November.
"It is important that we share with others in the grid community what we have learned, via concrete demonstrations of applications on our testbed," said Dr. Jysoo Lee, Deputy Chair of the PRAGMA Steering Committee, and Head of Supercomputing Research Department of KISTI.
In addition, the PRAGMA Steering Committee reviewed plans for the upcoming PRAGMA 5 Workshop to be hosted by NCHC in Hsinchu, selected the Computer Network Information Center (CNIC) of the Chinese Academy of Sciences as the host of the PRAGMA 6 Workshop on May 16-18, 2004 in Beijing, and added two new members, Kasetsart University and Academia Sinica Computing Centre (ASCC); these additional members will bring new expertise, applications, and resources to PRAGMA.
"We are very pleased to be able to include these two institutions within the PRAGMA family," said Peter Arzberger, Chair of the PRAGMA Steering Committee. "These two groups will contribute to realizing PRAGMA's dual mission of building sustained collaborations and to advance the use of grid technologies in applications."
"Collectively, the PRAGMA team has set out an ambitious set of plans for the next several months," said Dr. Arzberger in concluding the meeting. "By working together, we will be able to achieve our goals."
PRAGMA is supported by the National Science Foundation (Grant No. INT-0216895), the University of California, San Diego, the San Diego Supercomputer Center, The California Institute of Telecommunications and Information Technology, and member institutions.
For more information on PRAGMA, please visit http://www.pragma-grid.org/.