On the Influence of Start-Up Costs in Scheduling Divisible Loads on Bus Networks
IEEE Transactions on Parallel and Distributed Systems
Scheduling Divisible Loads in Parallel and Distributed Systems
Scheduling Divisible Loads in Parallel and Distributed Systems
RUMR: Robust Scheduling for Divisible Workloads
HPDC '03 Proceedings of the 12th IEEE International Symposium on High Performance Distributed Computing
A method for scheduling heterogeneous multi-installment systems
ACIIDS'11 Proceedings of the Third international conference on Intelligent information and database systems - Volume Part II
New method for scheduling heterogeneous multi-installment systems
Future Generation Computer Systems
| Hi-index | 0.00 |
Most common jobs of Grid computing are arbitrarily divisible. Divisible load theory(DLT) provides the mathematical machinery for time-optimal processing. With multiple round load distributions, idle processor periods can be harnessed for useful computation. Optimized rounds for the purpose can be planned in advance. The Grid is dynamic in nature. The above theory does not fully account for this. Any realistic scheduling strategy based on DLT has to take this fact into account. Existing multiple rounds algorithms do not involve time-varying effects due to environmental changes. This is a situation that leads to processing delays, such as Disk I/O contention. The proposed inter-round scheduling algorithm takes this into consideration. It involves time-varying resource performance degradation and results in resonable performance.