The Information Power Grid is changing our world

According to Gentzsch, the CERN Grid Challenges are enourmous. For new Large Hadron Collider project (LHC) CERN needs immensive data processing and storage cappacity (100 Tip/s). The petabytes of data generated have to be stored and made accessible for analysis. The structure is organised in three tiers. At CERN itself full data are available. In the second tier, at national level (called Regional Centres by CERN), serve as a focal point for national research and as a connection to the university departments that constitute the third tier.

This shows some resemblance to the power grid organisation. But Gentzsch pointed to another analogy, that of the "automotive power grid", although we use to call that "traffic infrastructure". In the automotive power grid, you have distributed objects (cars, buses, and the like), supporting technology, supporting infrastructure, like roads, lightning, services - like gas stations. After 200 years of development, today, this provides access to mobility, for anyone, anywhere, anytime, calims Gentzsch.

The computational grid shoud do the same, for computational capacities: It should be an infrastructure that couples computers, software, and other devices and make it avialbele over the Internet presented as a unified integrated (single) resource.

Why would we need the Grid in the first place? According to Gentzsch, we need it because it allows new applications made possible by the high-speed coupliing of people, computers, databases, intranets, and many more.

Before we can all enjoy this being emersed in computing all of the time, there are challenges to be met. Although part of the underlying technology has already been pursued in cluster computing and meta computing, the scale of the Grid brings in new problems, that have to do with bandwidth, interoperability, single global name space, resource management, to mention a few.

Gentzsch does not think bandwidth to be a problem in the future. Despite the problems most Internet users are facing each day, so much new fiber is put into the ground, that combined with new techniques to move more data over that fiber, we can assume Terabit/s networks to be abundant in the near future.

There are already many projects all around the world where has been experimented with Grid technology: DISCworld, EROPPA, JavaLine, Unicore, Wamm, just to mention a few. In EROPPA for instance, two post production companies, Mangos in Greece and Raster in The Netherlands, were connected to the supercomputer centres in Eindhoven and Athens. Render jobs from the post production companies could be automatically executed at the place where there was capacity available, be it locally, at the nearest supercomputer centre or at the centre at the other side of Europe. Genias Benelux delivered the gluing software.

Another emerging trend are the Application Service Providers. Although most of them provide access to business software, there are a few that also provide access to computational resources. According to Gentzsch, here we are just at the beginnning. A whole range of telecomputing, teledata and telelearning systems will emerge.

In fact, there seems no limit to future Grid scenario's. Application power from a wall socket, access to any resource anytime, any place; getting high quality healthcare independent on where you are. Gentzch concludes that whenever we are able to split up a complex and expensive task into a simple and cheaper client and a acomplex processing part, we will embed it into the Grid.