CluViz – A Web-Based Cluster Health Visualization Application
David FrankOral Defence Date:
Wednesday, October 17, 2007 - 15:00Location:
Professors Dragutin Petkovic, Bill Hsu & Researcher Mike Wong
The project, named CluViz, is a Web-Based Cluster Health Visualization system that graphically visualizes the functions of a cluster over a large period of time in order to monitor its health. CluViz comes under the category of System Management Applications (SMA), which employ visualization techniques showing the various aspects of the performance of the cluster. Cluster health can be defined as the proper functioning of the various components of the cluster and their collaborative behavior. Any deterioration in the cluster health could potentially slow or stop cluster performance. In this project, we designed, implemented and tested CluViz in the context of monitoring the Dell High Performance Cluster in San Francisco State University’s Center for Computing for Life Sciences. We also believe that the CluViz code and concepts can be extended to other systems. CluViz gathers the data related to the HPC’s health such as note and user information, and visualizes the data in various visualization modes. The User Interface of CluViz is made very simple for all cluster users in using the system to see the cluster health. Security in the CluViz system is enforced by having the application available only to users who have access to the cluster. We tested the system for correctness, performance and usability.