The Community for Technology Leaders
Cluster Computing and the Grid, IEEE International Symposium on (2009)
Shanghai, China
May 18, 2009 to May 21, 2009
ISBN: 978-0-7695-3622-4
pp: 268-275
ABSTRACT
We propose information aggregation as a method for summarizing the resource-related information, used by the task scheduler. Through this method the information of a set of resources can be uniformly represented, reducing at the same time the amount of information transferred in a Grid network. A number of techniques are described for aggregating the information of the resources belonging to a hierarchical Grid domain. This information includes the cpu and storage capacities at a site, the number of tasks queued, and other resource-related parameters. The quality of the aggregation scheme affects the efficiency of the scheduler’s decisions. We use as a metric of aggregation efficiency the Stretch Factor (SF), defined as the ratio of the task delay when the task is scheduled using complete resource information over the task delay when an aggregation scheme is used. The simulation experiments performed show that the proposed aggregation schemes achieve large information reduction, while enabling good task scheduling decisions as indicated by the SF achieved.
INDEX TERMS
Grid, monitoring, aggregation, hierarchical networks, scheduling
CITATION

E. Varvarigos and P. Kokkinos, "Resource Information Aggregation in Hierarchical Grid Networks," Cluster Computing and the Grid, IEEE International Symposium on(CCGRID), Shanghai, China, 2009, pp. 268-275.
doi:10.1109/CCGRID.2009.63
88 ms
(Ver 3.3 (11022016))