Workload Management Work Package
The workload management section of the DataGrid project has the goal of defining and implementing an architecture for distributed scheduling and resource management in a Grid environment. Because of the general-purpose nature of the GRID, this infrastructure implies an unpredictable, chaotic workload generated by relatively large numbers of independent users in contrast to the supercomputing environments that have been the target of most previous meta-computing projects.
Many challenging issues must be considered: optimizing the choice of execution location based on the availability of data, computation and network resources, the optimal co-allocation and advance reservation of CPU, data and network, the definition and implementation of uniform interfaces to different local resource management systems, the possibilities to define priorities and policies on resource usage, etc... This workload management system must be used in real production environments, so capabilities such as reliability, fault tolerance, scalability are very important.
A profitable collaboration with the Globus
Team and the Condor Team has been started since June 1999 with the aim of
understanding in depth the relevant aspects of the Workload Management
A profitable collaboration with the Globus Team and the Condor Team has been started since June 1999 with the aim of understanding in depth the relevant aspects of the Workload Management architecture.
DATAGRID Project Partners involved in Workload Management WP
Last update 11 Apr 2003