The Research On Grid Monitoring System Based On Web

With the improvement of Web technology, the web real-time monitoring systems are becoming a hot point and has already been implemented. With the friendly user interface, the grid monitoring system based on Web can obtain and operate parameters of grid resources which are distributed, large-scale, dynamic and heterogeneous. As the core function of the grid management and grid optimization, grid monitoring system helps the job management model to find the available services and resources, inform status of the object service to user. Furthermore, it helps managers of the resources to analyze the system's performance, find the problem and fix it, allocate the resources properly.In this paper, a browser/server architecture is proposed to construct grid monitoring system--CGSVWeb and a prototype system implementation is also presented. This architecture is well adaptive and scalable by using grid service invoking technique and adopting some features of hierarchical model and GMA architecture. The use of different information gathering mechanisms meets different monitor requirements and enables the flexibility and reliability in monitor system.The deployment of large scale grids generates new challenges in system and network monitoring. Users and distributed applications need to access logical and up to date information on sites, links performances and services availability. On an upper level, CGSVWeb has been designed to logically and graphically represent all elements, applications and services running over grids. It polls grid objects with different methods or make request to build aggregated views of grids on a Web Site, like logical views of services and full tree of computing elements. It provides the methods for predicting the performance of Computational Grid resources. The predictions generated in this manner are intended to support adaptive application scheduling in Grid settings, and on-line fault detection. CGSVWeb offers a flexible presentation layer in huge and heterogeneous environments, as it is in grid contexts. It provides a very simple, autonomous and extensible model that enables the visualization of any level of abstraction (virtual organizations, set of data stores, etc) needed by such environments.