This article will present the concept and implementation of a resource management system (RMS). The central component of the RMS is the resource broker GORBA that plans resource allocation by a combination of heuristic processes and evolutionary algorithms. For resource planning, schedules are generated, which distribute the grid jobs to the grid resources in a defined time window. A test environment with extensive visualisation options was developed for GORBA, which will be presented in detail. Using this test environment, benchmark runs were carried out, which are needed to evaluate and further develop GORBA. Automated resource planning and the graphic visualisation options facilitate the usability of a grid environment.
[1]
Olaf Schneider,et al.
The CampusGrid Test Bed at Forschungszentrum Karlsruhe
,
2005,
EGC.
[2]
Wilfried Jakob.
HyGLEAM - An Approach to Generally Applicable Hybridization of Evolutionary Algorithms
,
2002,
PPSN.
[3]
Achim Streit,et al.
Scheduling in HPC Resource Management Systems: Queuing vs. Planning
,
2003,
JSSPP.
[4]
Wilfried Jakob,et al.
Optimised Scheduling of Grid Resources Using Hybrid Evolutionary Algorithms
,
2005,
PPAM.
[5]
Christian Blume,et al.
GLEAM - An Evolutionary Algorithm for Planning and Control Based on Evolution Strategy
,
2002,
GECCO Late Breaking Papers.