Grids for Science - why all the excitement?Amsterdam 07 March 2001 Grids attract a lot of attention. Is it all buss or is there something real behind it. According to Paul Messina, who lectured at the Global Grid Forum in Amsterdam, Grids enable large-scale computational science. He used the Virtual Observatory that will include space data from many sources and make them accessible to all of the astronomical community, as an illustration.
Messina notes that today large computations on remote systems are feasible. We see harvesting of computing
cycles from thousands of network connected systems.
Squeezing more science out of investments in research facilities is always important. Through the Grid, the
vast amounts of scientific data that sensors, instruments and particles detectors will be made available.
Remember that research instruments, and satellites cost a great deal. Through the Grid data from these
devices cab be accessed and analysed by more scientists.
Also tele-instrumentation will be made easier by the Grid. Instruments can be controlled remotely: data
gathering and analysis can be done from a researcher's institute. For instance ARAN, the
Trans-pacific tele-microscopy collaboration.
What is new about this? Most of these types of experiments already have be done before. But according
to Messina, with the Grid they can now be carried out routinely, not as a stunt. Also the scale at
which these experiments can be performed is much larger and many can now be formed simultaneously.
The data explosion is another driving force of the Grid.
The Global Grid Forum is instrumental in this because of the coherence it gives to the community.
According to Messina, the Grid is made possible by a confluence of trends:
This new revolution will change the way research is carried out. Integration of data sources on the web is one example, such as NCMIR that federates brain data. An important one in astronomy is the Virtual Observatory that came out of the digital sky project. This is a federation of sky survey data from optical, infrared and radar instruments. Multi-terabyte of data are involved. The idea is to investigate a piece of the sky, combining data from all these different sources. That is why US National Virtual Observatory (NVO) was proposed. The NVO will archive data sets of space and supply a digital archive that can be accessed through a meta-data management tool. The NVO will serve the whole astronomical community, not only the US part. After the Virtual Observatory will become operational, astronomers will first have to do investigation on it, before getting time on a real telescope. Grids make it possible to link data, resources, instruments, to make this happen. Paul Messina envisages that the Grid will strengthen the position of simulation as the third method of doing Science (besides experiment and theory) and that it will expand the number of people that can do useful research with important data sets.
Ad Emmen |
 |