Resource management architecture for metacomputing systems

摘要

Metacomputing systems are intended to support remote and/or concurrent use of211u001egeographically distributed computational resources. Resource management in such 211u001esystems is complicated by five concerns that do not typically arise in other 211u001esituations: site autonomy and heterogeneous substrates at the resources, and 211u001eapplication requirements for policy extensibility, co-allocation, and online 211u001econtrol. We describe a resource management architecture that addresses these 211u001econcerns. This architecture distributes the resource management problem among 211u001edistinct local manager, resource broker, and resource co-allocator components and 211u001edefines an extensible resource specification language to exchange information 211u001eabout requirements. We describe how these techniques have been implemented in the 211u001econtext of the Globus metacomputing toolkit and used to implement a variety of 211u001edifferent resource management strategies. We report on our experiences applying 211u001eour techniques in a large testbed, GUSTO, incorporating 15 sites, 330 computers, 211u001eand 3600 processors.