With the help of GRID-computing it is possible to process huge amounts of data on computing farms all around the world, without the end-user has to know the individual properties of the the computing farms (i.e. version of operating system, security settings, number of CPUs etc.). Ideally a user submits a job, which is processed somewhere on the world and the output is copied back.

GRID-computing already plays a big role for experiments like DØ, CDF, Belle etc., as the produced data amount is barely processable by local computing farms. This situation will become even more critical after launching the experiments at the LHC. GRID-computing will than become mandatory.

At DØ a software package (SAMGrid) was programmed, which complies with the requirements and is in excessive usage. The whole Monte Carlo production is handled by SAMGrid nowadays. Additionaly it is possible to reprocess all recorded data within just a couple of weeks. The next step will be the dislocation of primary proceesing of the data through external computing farms by SAMGrid.

A big effort for development and certification of SAMGrid has been done by the german workgroups and they made a significant contribution to the success of GRID-computing at  DØ.

In Germany Monte Carlo production for DØ is done at the computing center in Karlsruhe (GridKa, 1500 CPUs).

Steering and controlling of the Monte Carlo production at GridKa has been taken over by the Mainz group.

Monte Carlo production at DØ with involved computing farms