A web services based Grid performance data management framework

Grid technologies have been widely adopted and are entering the mainstream in science and engineering computing. Grid technologies and infrastructures support the sharing and coordinated use of diverse resources that are geographically distributed components operated by distinct organizations and individuals with different policies. In a Grid environment, a substantial amount of runtime information requires to be gathered and delivered for various Grid services such as performance analysis, performance tuning, performance prediction, fault detection, co-scheduling and re-scheduling. Because the characteristics of runtime Grid performance data are fundamentally different from those of system or program-produced data, systems that collect and distribute the information should satisfy certain requirements. Exploiting Web services technologies, we persent a Web services based Grid performance data management framework to fit with the heterogeneous, dynamic and multi-domain nature of Grids. In order to meet the requirements for QoS, some security mechanisms are incorporated into the framework.