dc.description.abstract | ‘Grid computing’ is a recently coined terminology that captures the desire to offer
computing services in a widespread, transparent and reliable manner similar to the
way electrical grid systems supply electricity. Grid computers consist, respectively, of
several computing elements such as processors, clusters, personal computers, data
storage facilities and other peripherals that appear to the users as a single high
performance computers. The usage is transparent in the sense that the users do not
have to know where the resources are located.
Since Grid computers consist of several interconnected computing devices, matching
users and resources still remains an interesting challenge. This challenge is more
demanding because the resources that are shared span across geographically
independent administrative domains with different policies. The resources may appear
or disappear from the grid computing pool without notice.
In this paper, we review the scheduling strategies used in grid computing and identify
multi-agent systems based scheduling options. We present our model that investigates
how multi-agent systems can be used to support scheduling functions in a dynamic
environment. Our main contribution is a more focused isolation of scheduling
strategies in grid computing and their four-group categorization. We also examine the
feasibility of setting up a campus grid computing services in an institution of higher
learning such as a University in a developing country. | en |