基于性能量化矩阵的计算网格作业调度算法研究

Research on Job Scheduling Algorithm of Computing Grid Based on Performance Measured-Matrix

提升计算网格系统运行效率的关键在于作业调度算法,如何综合各种因素使得调度策略更为全面是一个有挑战性的问题。通过建立网格资源性能量化矩阵,构建了一个作业调度模型,并基于此模型给出了一个具体的作业调度算法。通过性能分析和实验仿真,该算法在运行时间、占用资源等方面都有较大的改善,能较好地适应网格系统的动态性和可扩展性特点,提高了作业调度效率。

The key technology to improve the grid system operating efficiency is the job scheduling algorithm, how to integrate various factors to make the scheduling algorithm more comprehensive is a challenging problem. By establishing grid resources performance measured-matrix, the paper constructed a job scheduling model, and gave a specific job scheduling algorithm based on this model. Through the performance analysis and the experimental simulation, this algorithm improved the executing time, took few resources, and better adapted to the dynamic and scalability characteristics, which enhanced the job scheduling efficiency.