The Grid is about connecting peopleParis 07 March 2002 Is the Grid about connecting computers? Yes it is. Is the Grid about making very large amounts of data available over the Internet? Yes it is. Is it about connecting information sources to create knwowledge? Yes it is. And the list continues. The Grid is not just about very large scale efforts, but also about departments clustering, and about commercial nonstop-self repairing computers. And this is only a selection of the interpretations of what belongs to Grid computing that was heard on the DataGrid conference which was held in Paris. One interpretation, or definiton perhaps, caught my attention: the Grid is about connecting people. The whole DataGrid effort for instance allows 6000 physicists to collaborate better, by having access to the same data, the same programmes and a collaborative working environment. This vision was expressed by Manuel Delfino from CERN during a panel discussion on March 7.
Indeed, when one thinks about it, this definition makes sense. The Grid is then a collaborative environment, where complexities of moving data, running programmes and user management issues are incorporated in the system and invisible to the user. The Grid is about infrastructure. The reason why this is happening, according to Manuel Delfino from CERN who quoted a mid 2000 report, is that large teams of highly qualified creative people are working better when they are not all put in one place. Hence a tendency for larger companies to distribute that research and development teams, but let them collaborate closely. This, in fact resembles the situation in European high-energy physics research where there are many institutes all over Europe collaborating on a few large experiments.e As Manuel Delfino said during the panel discussion, the research network in Europe (Rolled out in the European Géant project) is the best in the world. Also a lot of high-quality research is there. So there is a fertile ground for research, applications and even businesses. Looking at the discussion panel composition, one sees however, that in the chain, from raw research, one link is missing. There are universities and research institutes. And there are system integrators (like CS, DATAMAT, and EADS). All companies that were on the panel selling Grid building blocks, be it integrated hard and software, or software like resource mangement software were American. For instance IBM, SUN, HP, and Platform computing. This is indeed typical for the situation in Europe. If we look at the EnterTheGrid.com catalogue there are only a few companies (although some are very active) here providing building blocks for the infrastructure that the Grid is. In this way the situation is different from the beginning of the supercomputer era, when there were several European efforts to build a supercomputer. In the end they all failed... Just as the Web was not possible before TCP/IP was there, Ian Foster said during the panel discussion the Grid was made possible by Linux, which paved the way by experimenting with open source and the licenses for it, that were both acceptable for research and for industry. Finally, Mr. Spire from the French newspaper Le Monde who chaired the round table said in his concluding remarks that as far as the Grid is concerned, there is starting to emerge a transfer from research to industry. At present research is asking questions to industry. Research has been developing the Grid concept and now turns to the industry for collaboration. Mr. Spire was happy to see this transfer taking place at the round table discussion, where there was a lively discussion between the researchers and the representatives from industry.
Ad Emmen |
 |