Optimizing locationing of multiple masters for master-worker grid applications

Abstract

The problem of allocating a large number of independent tasks to a heterogeneous computing platform is considered. A non oriented graph is used to model a Grid, where resources can have different speeds of computation and communication. We claim that the use of multiple masters is necessary to achieve good performance on large-scale platforms. The problem considered is to find the most profitable master locations in order to optimize the platform throughput. © Springer-Verlag Berlin Heidelberg 2006.