Comparison of pricing policies for a computational grid market

Abstract

In this paper, we demonstrate and discuss the economical results of applying a fixed, a dynamic and a stochastic approximation based pricing policy in a free commodity market model designed for computational grids. We present the pricing policies and our economy-driven scheduling heuristic as a part of our market model in which we assume resource owners desire to profit or recover their costs, and users desire to execute their jobs within the limits of their budget and time. The comparison experiments revealed that our dynamic pricing policy is more successful as a means for achieving social welfare in the market.