Performance-based middleware services for grid computing

Managing resources and distributed workload across multiple administrative boundaries is a key issue in grid computing and middleware research. Standard services that perform look-up, match-making, scheduling and staging are being developed to address this problem. As active middleware components, these services will allow users to identify and utilize appropriate resources that provide sustainable system- and user-level qualities of service. This paper documents two enhanced match-making services that address the performance implications of executing a particular workload on a given set of resources. These services are based on an established performance prediction system that is employed at both the local (intra-domain) and global (multi-domain) levels to provide dynamic workload steering. These additional facilities bring about significant performance improvements, the details of which are presented with regard to the user-perceived quality of service and to the grid resource utilization.