期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2008
卷号：8
期号：6
页码：42-48
出版社：International Journal of Computer Science and Network Security
摘要：Clusters have emerged as a primary and cost-effective infrastructure for parallel applications, including communication-intensive applications that transfer a large amount of data among nodes of a cluster via interconnection networks. Efficient resource usage is a key to achieving better performance in cluster systems. Previously, most research in this area has focused on balancing the load of each node to use the resources of an entire system more effectively. Memory utilization has a key role in the implementation of a high performance cluster. However determining the load without considering the memory utilization will degrade the cluster performance. In this paper, a load metric has been proposed, termed as Tendensive Weight, in order to balance the cluster dynamically. This metric considers both CPU and IO utilization including memory exploitation. Simulation results shows that the system incurs 10% shorter execution time than the conventional approach using number of effective tasks.
关键词：HPC, Load Balance, CPU utilization, Memory exploitation, Tendensive Weight
