Design And Implementation Of Cluster Monitor And Management System

Cluster system technology is applied more and more extensively either in production or in daily life. The way to solve computational problem of complex model and handle massive amounts of data, using the parallel computation ability of cluster system which has the low cost and high performance characteristic, has become important direction in high performance computing field.The quality of cluster resource monitor and management system directly affects the performance of cluster system. The traditional way of cluster monitor and management by command line is complicated. Manager of cluster system has to remember a lot of commands for monitoring and managing the cluster system. However, in recently, with the development of B/S modal, more and more cluster management systems can be used based on web application. It is convenience for user to operate cluster system through web interface. What’s more, the information of cluster system is presented in the form of image. This paper is aimed at designing and implementing a cluster monitor and management System which is based on web application.During the project, the author participates in the work of requirement analysis, preliminary design and detailed design with others. After these works, the author implements the functions of the system by himself.Understanding the conception and application about cluster system, it’s helpful for designing the system. By investigation of cluster system’s technology, the author analyzes the architecture and principle of the cluster system, includes the key contents that necessary to be done in the procedure of cluster monitoring and managing work. Based on the research, this paper introduces requirements analysis and preliminary design of the system, in which build function modal and data modal with applying the software engineering method. All of these are aim at supporting detailed design and implementation.The system implemented follows the MVC software design principle and developed with SpeedPHP framework. As the basic layer, Ganglia and TORQUE are integrated into the web application through their command and API.It’s improved by testing that the system which introduced in this paper achieves the anticipated demand and reduces costs of learning how to use and maintain the cluster system. According the result of using, the system has good application prospect.