Grid Resource Management Information Services for Scientific Computing
Lim Choi Keung, H.N., Wang, L., Spooner, D.P., Jarvis, S.A., Jie, W. and Nudd, G.R. (2002) Grid Resource Management Information Services for Scientific Computing. In: Proceedings of the International Conference on Scientific and Engineering Computation (IC-SEC 2002), 3-5 December 2002, Singapore.Full text not available from this repository.
Scientific applications typically have considerable memory and processor requirements. Nevertheless, even with today's fastest supercomputers, the power of available resources falls short of the demands of large, complex simulations and analysis. The solution fortunately lies in the emergence of innovative resource environments such as Grid computing. This paper addresses issues concerning the discovery and monitoring of resources across multiple administrative domains, which can be harnessed by scientific applications.
Grid Information Services are an important infrastructure in computational Grid environments as they provide important information about the state and availability of resources. We will present how the Globus Monitoring and Discovery Service (MDS) architecture can be extended to offer a transparent, unified view of Grid resources. Local resource managers ultimately possess the potential to influence where scientific jobs are processed, depending on the availability and load of their scheduler. Consequently, information about the structure and the state of schedulers should be provided to Grid brokers, which will then decide where to schedule the applications.
This paper will present the way in which low-level scheduling information is collated from multiple sources and is incorporated into the unified Grid information view. The existence and availability of Grid resource information also allows reliable application performance prediction to be carried out using the Warwick PACE (Performance Analysis and Characterisation Environment) system. Thus, Grid Information Services are a key component in the efficient execution of scientific applications on Grid computing architectures.
|Item Type:||Conference or Workshop Item (Paper)|
|Uncontrolled Keywords:||pcav hpsg grid middleware management performance|
|Subjects:||Q Science > QA Mathematics > QA75 Electronic computers. Computer science|
|Divisions:||Faculty of Science > Computer Science|
|Depositing User:||Simon Hammond|
|Date Deposited:||02 May 2011 18:34|
|Last Modified:||23 Feb 2012 09:08|
Actions (login required)