摘要
随着数据规模的不断增大,各个领域都需要海量高性能计算能力。网格计算效率的提高关键在于合理安排任务调度。论文在分析异态最早完成时间(HEFT)和可靠性动态水平调度(RDLS)算法的基础上,借助HEFT的优势对RDLS算法进行改进(HRDS),包括本地和全局可靠性驱动调度。结果表明HRDS算法性能明显优于HEFT算法,而且具有很好的可靠性和效率。
With the increasing scale of data ,all fields require a lot of high-performance scientific computing .The key of Grid computing efficiency lies on reasonable arrangements for task scheduling .This paper analyzes heterogeneous earlist fin-ish time(HEFT) and the reliability dynamic level scheduling(RDLS) algorithm ,an improved RDLS(HRDS) is proposed based on the advantages of using HEFT algorithm ,and it is divided into local and global reliability driven scheduling .Sim-Grid simulation results show HRDS algorithm outperforms HEFT algorithm ,and has good reliability and efficiency .
出处
《计算机与数字工程》
2014年第6期927-929,958,共4页
Computer & Digital Engineering
基金
江苏省自然科学基金项目(编号:BK2011319)资助