Get connected - welcome to the grid

The sky’s the limit as the next incarnation of the web brings data-processing powers only dreamt of until now. Anna Fazackerley looks at how projects such as a virtual observatory for astronomers will benefit

Oxford physics graduate Tim Berners-Lee invented the worldwide web in 1989. Now few academics can imagine life without it. But, behind the scenes, scientists are working on the next generation of the web, called the grid.

Just like an electricity grid, the “supercomputer” network will supply power on demand: scientists constrained in their work by their computer’s memory, software and processing power will be able to run programs all over the world at the press of a button. The grid is already big business in research terms, having attracted £213 million from the Office of Science and Technology.

One of the most potent examples of how the grid will help science is the large hadron collider, a massive nuclear collision experiment being constructed at Cern, the European particle physics laboratory in Geneva. The particles fired down its length, to simulate the Big Bang that created the universe, will generate enough data in a year to fill a stack of CDs the height of the Eiffel Tower. Without the grid, dealing with this data would be unthinkable.

Particle physics is not the only area to experience data problems. Andy Lawrence of the Institute for Astronomy at Edinburgh University is heading AstroGrid, a £3.7 million project to create a virtual observatory for astronomers in the UK. Lawrence says that his astronomy research naturally led him to the grid. “I’m involved in the big sky survey, so a couple of years ago I was wondering how we would cope with all this data,” he says.

He explains that data sets in astronomy are growing all the time and the new instruments being developed will only intensify the situation. Sky surveys previously took place once every few decades. In 2006, when Vista (the visible and infrared survey telescope for astronomy based in Chile) comes online, sky surveys will take place every few days and produce hundreds of terabytes of data.

Sheer volume of data is one problem, the variety of sources another. AstroGrid project scientist Nic Walton, of Cambridge University, says: “Any high-profile astronomy problem involves looking at the sky in different ways. We need to acquire and assimilate data from many different sources.”

Speed is vital. The large telescopes are in remote locations to get the clearest view of the sky, but researchers working on the other side of the world ideally need instant access to the data. Walton says: “If a gamma ray burst goes off, in seconds you want to be providing information about its location, so that other ground-based facilities can start observing, since they fade in the space of hours.”

So how do astronomers access and use all this information? Lawrence says that lots of data are available online, but the scientist needs to know exactly where to locate them. Also, even when the data sets are retrieved, they are often in different formats or not accessible to all. AstroGrid and other similar ventures should change that. The International Virtual Observatory Association will ensure that all the astronomy databases speak the same language. Researchers want the databases to be as simple as the internet.

“The beautiful thing about the web is you type in an address and the page comes up. You can see pages across the world and it feels as though they are in your computer,” Lawrence says. The aim of the virtual observatory is to achieve the same feeling with archives and data.

Lawrence’s team may be enmeshed in the technicalities of the grid, but it does not want the end user to consider them. “The idea is that scientists will get their hands on the data without thinking about how the data were acquired,” Walton says. “It will be like a Google search: that involves a massive amount of technology below the surface, but as a user you don’t need to know.”

The virtual observatory looks set to change the face of astronomy. “We will ultimately end up with more science, more astronomy and more papers being published,” Lawrence says.

Walton hopes that astronomy can lead the way for other disciplines that have yet to harness e-science.

In reality, it may be some time before the grid revolution really takes off. Project manager Tony Linde, a developer brought in from outside academia to give AstroGrid a commercial dimension, says: “In the wider world there’s not much awareness of the grid, and in academia there is not much except among those who are directly involved.”

But the AstroGrid team is not letting this deter it. The project will be completed at the end of next year, and the team already plans to seek funds for a follow-up. Lawrence concludes: “Whatever the fashions are, we don’t want to just have clever ideas, we want to really finish something that works.”

ICT in Higher Education No.2 index page