Nature-Inspired Algorithms Applied to an Efficient and Self-adaptive Resources Selection Model for Grid Applications

Autor

María Botón-Fernández, Francisco Prieto Castrillo y Miguel A. Vega-Rodríguez

Capítulo de libro
Theory and Practice of Natural Computing, LNCS, Vol. 7505.
Springer-Verlag, Berlin Heidelberg, Germany, 2012, pp. 84-96.
ISBN: 978-3-642-33859-5

Abstract

Grid computing infrastructures are systems composed by an heterogeneous and geographically distributed resource set. Despite the advantages of such paradigm, several challenges related to grid resources selection and resources availability still demand active research.

The aim of this article is to provide an efficient and self-adaptive resources selection strategy for grid applications deployment. This re- sources adaptation capability is provided by applying nature-inspired algorithms during the selection process. Specifically, both the preferential attachment technique from Complex Network field and a cellular automata model are used.

Finally, the results obtained during tests in a real grid show that the proposed model achieves an effective use of grid resources, resulting in a reduction of application execution time and in an increased rate of successfully finished tasks. In conclusion, the model improves the infrastructure throughput for grid applications.