An Economy Driven Resource Allocation Middleware for Grid Workflow
In order to accomplish high-performance on grid workflow, grid resource management system needs a smart and swift resource allocation middleware. In this paper, the authors study the economy driven resource allocation problem based on market model of grid resource management architectures. They model the problem as the Multiple Choice Knapsack Problem (MCKP) and design the resource allocation optimization algorithm to minimize the average turnaround time of the grid workflow. The complexity analysis shows that the optimization algorithm leads to more efficient resource allocation than many current algorithms. The algorithm is also proved to be especially appropriate since the problem's own characteristics can accelerate the algorithm implementation process.