Grid enabling 'real' science and engineering

Abstract

The Grid was proposed in the late 1990's to enable the next generation of research in science and engineering. The idea was forward looking and bold, but in some ways remarkably simple. It involved connecting high performance computers, large data servers and scientific instruments by high speed networks to allow large scale experiments and computations to be performed both seamlessly and routinely. A critical component in this vision was the development of standard 'middleware', which has taken significant effort and been the focus of large international research and development teams. Ten years later, we are starting to see sufficient maturity in the software systems that it is now possible to do 'real' science.

In this talk I will discuss our experience in building Grid testbeds, and in building robust software tools for performing scientific experiments. In particular, I will highlight a number of successful projects in environmental science, systems biology, chemistry and physics.