An Efficient and Self-adaptive Model Based on Scatter Search: Solving the Grid Resources Selection Problem

Autor

Francisco Prieto-Castrillo, Antonio Astillero, Elena Rey-Espada, María Botón-Fernández, María Arsuaga-Ríos

Capítulo de libro
R. Moreno-Díaz et al. (Eds.): EUROCAST 2013, Part I, LNCS Vol. 8111.
Springer-Verlag, Berlin Heidelberg, Germany, 2013, pp. 396-403.
ISBN: 978-3-642-53855-1

Abstract

Grid computing environments are distributed systems formed by a heterogeneous and geographically distributed resource set. In spite of the advantages of such paradigm, several problems related to resources availability and resources selection have become a challenge extensively studied by the grid community in last years.

The aim of this work is to provide an intelligent and self-adaptive model for selecting grid resources during applications execution. This adaptive capability is obtained by applying during the selection process an evolutionary method known as Scatter Search (it is based on quality and diversity criteria).

Finally, the model is evaluated in a real grid infrastructure. The results show that the infrastructure throughput is enhanced. Even more, a reduction in the applications execution time and an improvement of the successfully finished tasks rate are also achieved. As a conclusion, the proposed model is a feasible solution for grid applications.