Matchmaking: Distributed Resource Management for High Throughput Computing
HPDC '98 Proceedings of the 7th IEEE International Symposium on High Performance Distributed Computing
| Hi-index | 0.00 |
A decentralised resource management technique is described that makes use of 'resource capabilities' and 'task requirements' to find suitable allocations. The technique extends the 'classified advertisement' approach adopted within the CONDOR [3] system, to include resource usage history and variable resource granularity. Our approach can deal with a dynamically changing environment, which could include new tasks created at run time, or new devices added to existing clusters. A managed cluster can consist of identical or heterogeneous resources, with usage dictated by a local usage policy. The approach makes use of various algorithms that use similarities between devices to find approximate matches. Each resource administrator is required to describe the resource using a policy description scheme, which is subsequently used by a MatchMaker to find suitable allocations.