A novel price prediction scheme of grid resources based on time series analysis

Abstract

Grid computing generally allows heterogeneous resources to achieve resources sharing and service-oriented scheduling. Grid economy-based management has been viewed as a feasible approach to carrying out fair, efficient and reliable scheduling. One key issue in grid economy-based strategy is to contain accurate and predictable prices of heterogeneous resources. In this paper, we present a novel price prediction scheme of grid resources for achieving accurate prices for resource management in the context of grid economy. Our proposed scheme utilizes Time Series Analysis (TSA) tool to predict and generate resources prices. We examine the practical performance by presenting simulation-based results, which demonstrate that proposed scheme is feasible and efficient in the grid economy. © Springer-Verlag Berlin Heidelberg 2007.