GRID '02 Proceedings of the Third International Workshop on Grid Computing
On QoS-Based Scheduling of a Meta-Task with Multiple QoS Demands in Heterogeneous Computing
IPDPS '02 Proceedings of the 16th International Parallel and Distributed Processing Symposium
Scheduling of a meta-task with QoS requirements in heterogeneous computing systems
Journal of Parallel and Distributed Computing
Multiagent and Grid Systems - Engineering Environments in Multiagent Systems
| Hi-index | 0.00 |
An increasing number of applications execute over a set of computing and communication resources and have end-to-end quality of service (QoS) requirements. Given a system and an application with specific flow structure and QoS requirements, this paper describes an integrated QoS-based allocation and routing algorithm that determines which computing and communication resources to utilize for this application. More specifically, the algorithm finds the best flow path for an application, subject to application flow and QoS constraints, so as to optimize system objectives (e.g., minimize total system utilization or balance the load). The paper also introduces a delta-optimal algorithm that, relative to the optimal algorithm, significantly reduces the run-time of the algorithm with a minor degradation in the optimality of the solution. The delta-optimal algorithm is guaranteed to produce a result that is at most delta-units sub-optimal. The paper also analyzes the optimality and run time of the solution as a function of delta for two multimedia scenarios. Rules of thumb for setting delta are also given.