What is the Grid?Grid at workGrid ChallengeBref historique de la grilleLa grille et vousGrid at CERNProject in the world
Demos and Games
The scenario: a flood in Matagalpa
Imagine...

...that you are working at the UNOSAT office in Geneva. One morning while sipping your coffee and reading the morning paper, you suddenly receive a message from a disaster relief team in Matagalpa reporting a major flood is occurring in the valley following torrential rains, and that they require detailed satellite maps of the area as quickly as possible. You respond quickly and send an urgent request for satellite images of Matagalpa.

Satellites traveling in orbit between 650 and 850 km above the Earth respond to your request and map out the region of Matagalpa using sophisticated radar and optical techniques. The optical images show the topography of the countryside, while radar reflects very differently from water and land, and thus shows up the flooded areas clearly. The satellites then send the raw data (telemetry) back to Earth.

When the Ground Station receives the raw data from the satellite, it does some advance data processing on the data which yields several satellite images, each roughly 400 MB in size. These image files are then promptly relayed back to your office in Geneva via the Internet.

You put your long experience to practice and quickly analyse the 400 MB files, before you layer and combine the optical and radar images to increase the resolution of the pictures. You then "stitch" together the new image files to form a "mosaic" that maps out the full area affected by the flood. Using GIS tools (Geographical Information Systems), you highlight the local road networks, the villages and the zones most a risk on the mosaic. You are now ready to send the new image file (the mosaic), to the disaster relief team in Matagalpa without further delay.

But you have one problem...

Your new image file is now several Gigabytes large and must be compressed before you can send it to Matagalpa via a low bandwidth modem link. On your local computer, this is a time-consuming process that delays the relay of the mosaic to the needing aid worker by several hours (not compressing would, however, delay the transfer even more).

Time is of essence and you realise that you need to come up with a smart way around this problem. This is where the Grid can make a big difference: you don't have the extra resources you need, but you know that if you submit your job to the Grid, then you can instead make use of free computing power somewhere else.

What the demo shows is how this would work in practice. Basically, the idea is that replicas of the images you have stitched together are stored on several computers around the world. When you ask the Grid to do the compression, the sites where there is the most computing capacity immediately available to do the compression are prioritized, and the compressed tiles are sent individually, from wherever they are, to Matagalpa, where they are expanded and restitched on a local computer.

The few hours saved could mean many lives saved as well. Of course, there are many other aspects of this complex chain of events where increases of efficiency could be achieved by improved IT solutions. Grid technology is only part of a better solution. Still, this is probably a realistic assessment of the way Grid technology will gradually begin to impact the lives of ordinary people - even in developing countries.