HPC Cluster Monitoring System Architecture Design and Implement

摘要

High performance computing (HPC) facilities such as HPC clusters, as building blocks of Grid computing, are playing an important role in computational Grid. HPC monitoring in HPC cluster systems presents an important challenge because HPC cluster environments are volatile, heterogeneous, not reliable and are managed by different middleware and systems. In this paper, we propose an HPC cluster monitoring system with four tier structure for Grid computing and utility computing clusters. It provides the basic function such as job monitoring, and system monitoring, etc. With our prototype, the Grid users are able to find the available cluster nodes, and customize their preferred HPC cluster nodes for their computation intensive applications in Grid computing or utility computing.. Experiments show that our work provide great convenience and flexibility for users to make good use of HPC cluster.