Grid Computing: Starting for an e-ScienceMunich 13 January 2003 Dr. Marcel Kunze, Department Grid-Computing and e-Science, Research Center Karlsruhe described the situation at the Research Center, which will be part of the DataGrid project from CERN. He presented the requirements of the Grid Computing Competence Centre GridKa.
First he described the road from 1980 to 2005, starting with networking, TCP/IP, communications, e-mail, information, World Wide Web, Linux and now Grid Computing. He views it as a mobile, desktop and visualisation access through a grid middleware to supercomputers, PC-clusters, data storage, sensors, experiments, and Internet, networks. The GRID computing competence centre, GridKa, is a computer centre for GRID computing and a competence centre for e-science and concerted research.
He presented the increase in computing power of the centre. The power is based on Intel Pentium III, 1 GHz, starting with 25 in 2001, to 350 in 2002, 725 in 2003, 1675 in 2004, to 3875 in 2005, 8300 in 2006 and 22900 in 2007. The same will happen with the disk and tape storage, in 2003 there will be 113/211 TeraByte, which will increase to 1.421/3.737 PetaByte in 2007.
E-Science and Concerted Research
Both topics are more than only the sum of networks, GRIDs and High-Performance Computing. The e-Science is engaged in the global co-operation in key areas of research and with the next generation of infrastructure that allows this, as Dr. John Taylor, member of the British Science Council stated. "Concerted Research is our programme for the future, networking as a principle of researching thinking and doing." Dr. Kunze mentioned the Institutes and the Research Centres of the Helmholtz Gesellschaft, which are distributed all over Germany. They have a network connection of 2.5 Gbit/s, they want 10 to 100 Gbit/s.
One example was the Earth Observation, the ESA mission delivers 100 GByte (ERS 1/2) and 500 GByte (ENVISAT) daily. The requirements of the grid are coupling with complex sensors, reprocessing huge historical archives and complex work flows (datafusion, data mining, modelling, ..).
The Semantic Grid goes from e-science via the knowledge grid to the information grid and ends with the computing and data grid with its automatics - autonomic computing. This way he calls management and control. The way back from automatics to e-science is named from data to knowledge.
The European Commission funds Grid infrastructures. Dr. Kunze listed the future perspectives: extended application of the grid technology to every day's world, e-society, Cyber Infrastructure. He expects an international Grid infrastructure which hides the complexity for the user, invisible computing. The Grid computing one day will be an obviousness like the electric current. On the networks the Grid infrastructure is built, on which the grid-enabled applications are sitting.
Uwe Harms [News on Advanced IT] [Calendar] [Analysis] [IT in Medicine] |
 |