Algorithmic approach to calculating minimal resource allocation recommender for grid using reliability and trust computations

Abstract

In this paper we aim at minimally allocating resources on the grid by providing a recommender that gives us a recommendation of which allocation is a minimal resource allocation out of allocation matrices space. The recommender uses reliability, trust and search(2-way split backward search) algorithms. The resource allocation matrices are generated according to the search algorithm. The reliability and trust computations are used as factors to judge the minimal resource allocation from the resource allocation matrices generated by the 2-way split backward search algorithm [6]. Reliability is based on rate of failure of grid nodes, communication links and also on reliability of services on the grid. Trust is internally based on reputation which is overall grid service reliability and direct trust which is the availability of resources. In this work it is assumed that the arrival rate of failure of grid elements follows the Poisson process. © 2011 Springer-Verlag.