Task-Oriented Computational Economic-Based Distributed Resource Allocation Mechanisms for Computational Grids

In computational grids, heterogeneous resources with different ownerships are dynamically available and distributed geographically. It is not realistic to build the resource allocation mechanisms for such computational platform without considering economic issues. Developing computational economic-based approaches is a promising avenue for building efficient, scalable and stable resource allocation mechanisms without a centralized controller for computational grids. The key difficulty in building a computational economic-based resource allocation mechanism is measuring the economic value of resource usage. In this paper, we propose a task-oriented mechanism for measuring the economic value of using heterogeneous resources in computational grids. This mechanism provides feasibility for resource users to evaluate their outsourcing decisions. It also gives resource suppliers incentive to provide their resources to computational grids. Based on this mechanism, auction-based, commodities market-based and game theory-based distributed resource allocation mechanisms are established for computational grids. 1