Freedom, GEABSOLUTE POWERS CORRUPT ABSOLUTELY, General Election (GE15), Malaysia, Politics, polling Nov 19: Destroy Umno for the betterment of Malaysia, race, religion, Solidality, support Aliran for Justice

Share This

Saturday 16 January 2010

Intercontinental Grid Computing: Europe and China Link Up for Research

Intercontinental Grid Computing: Europe and China Link Up for Research

ScienceDaily (Jan. 15, 2010) — Grid computing can jet-propel research and development. An EU-funded programme that lets European and Chinese grids work together has already produced results in aircraft design, drug development and weather prediction.

In 2007, the EU-funded project, BRIDGE (for Bilateral Research and Industrial development enhancing and integrating Grid Enabled technologies -- see: http://www.bridge-grid.eu/), set out to link European and Chinese computing grids and enable researchers to carry out joint research.

The project was inspired by the realisation that China is rapidly becoming a world leader in research and development, as well as a booming market for European products. Developing the infrastructure to link computing grids was seen as a key step towards future scientific and industrial cooperation.

"If Europe does not want to lose ground, the response can only be to synchronise with these developments," says Gilbert Kalb, BRIDGE project coordinator.

Building a shared infrastructure

The BRIDGE team's first challenge was to make the software systems that manage the European and Chinese grids compatible. The European Grid infrastructure, GRIA, and the Chinese system, CNGrid GOS, provide comparable services, but are organised differently.

The team were able to get GRIA and GOS to work together by building a new software superstructure to access them and tap their capabilities. The system included new gateways into the two grids, plus a shared platform to manage overall workflow, access needed applications, and translate higher-level commands into steps that each grid could carry out.

Not surprisingly, security was an important consideration on both sides. Kalb says that many or the scientific and industrial problems that BRIDGE was developed to address require intensive cooperation, yet involve highly sensitive information.

BRIDGE resolved this issue by letting selected processes remain private. That allows one group to contribute data or results to all collaborating parties without having to share proprietary software or analytic tools.

"You can interface in terms of the input and the output, while the algorithms remain hidden," says Kalb.

Putting BRIDGE to work

The BRIDGE team tested the intercontinental grid they built by attacking three problems, each of which made different demands on the system.

Discovering new drugs remains an extremely costly process. One way to speed research is to use computers to simulate the chemical fit between millions of small molecules and proteins that play vital roles in disease-causing organisms. A molecule that binds strongly to a key protein has the potential to be turned into a potent new drug. This kind of research demands enormous computing power.

Researchers in Europe and China contributed four different docking tools -- programs that calculate bonding between a small molecule and a particular protein. Each program used a different approach and produced somewhat different results.

The researchers then examined millions of molecules to see if they held promise against malaria or the H5N1 bird flu virus. By combining the results of the four different simulations, they were able to identify promising molecules more efficiently.

"Making the outcomes of these different docking tools comparable is very new," says Kalb.

The four-pronged approach produced promising results. The BRIDGE infrastructure has already been adopted in Egypt to target the malaria parasite.

BRIDGE was also used to solve a complex aeronautic problem -- designing and positioning wing flaps to maximise lift and minimise noise as an aircraft lands.

Like drug-discovery, these aerodynamic simulations required huge computational resources. In addition, because different parts of each simulation took place in different research centres, optimising the flow of work from centre to centre was also challenging.

The BRIDGE team was able to meet these challenges, carry out intensive distributed computations, and determine optimal wing flap parameters. "It proved to be an effective method for solving multi-objective and multi-disciplinary optimisation in aircraft design," Kalb says.

Weather data on the fly

Weather and climate represent a third area where international cooperation is vital. The BRIDGE researchers set out to link three large meteorological databases located in Europe, North America and Asia.

The key challenge they faced with this project was to handle enormous volumes of data efficiently.

"You could do a calculation in the United States and transfer the results to Europe, or you could fetch the data from the USA and do the calculations here," says Kalb. "The best way to do it depends on what calculation and what data and what's the best available way to transfer the data from place to place. Bridge does all this on the fly."

"Because there was a big organisation behind it, and our work fits very well, it was taken up right away," says Kalb. "I believe that meteorologists are already using it to access data and perform certain calculations."

To Kalb, the importance of what BRIDGE accomplished goes far beyond any single piece of research. He feels that the project has built the foundation for the kind of multinational collaboration that is needed to tackle global problems.

"Problems like energy and climate change can only be attacked or really solved with efforts from different players around the world, and we've built a platform to do that," he says. We proved that this is feasible and useful. Now it's time for other people to jump on this, develop it further, and use it."

The BRIDGE project received funding from the Sixth Framework Programme for research.

1 comment: