Topic leader
Ramin Yahyapour, ramin.yahyapour@udo.edu
Short description
The existence of a flexible but highly efficient scheduling system is still a key component for any cluster, Grid, SOA or Cloud system. While there already exist many scheduling systems, the area of scheduling remains a basic research area with major impact for the proliferation of large scale distributed computing. The requirements for such scheduling systems gained significantly by added complexity of the particular application use cases. While historically scheduling was mostly focused on batch-oriented parallel or sequential jobs for parallel systems, we additional requirements as for instance:
- Jobs with dependencies as in workflows
- Multi-resource planning which combines compute resources with other resource types like licenses or data
- Multi-criteria optimization goals which consider not just execution time but other criteria like cost, power consumption, or minimization of resource foot print
A link to the CoreGRID Institute on RMS might be appropriate as the
research topic will refocus on the pre-existing work from the Institute
agenda:
/mambo/content/blogcategory/16/295/
Keywords
Scheduling, resource management, multi-criteria optimization, planning.
News
[Apr. 09] The CoreGRID RMS Institute meeting was be held on April 21/22nd in Pisa, Italy.
|
Last Updated ( Tuesday, 28 April 2009 )
|