In order to solve the problem of significant fluctuation of workload at each node, a new approach is proposed for acquisition of load parameter for Now, a network of heterogeneous computational resource. The current m...In order to solve the problem of significant fluctuation of workload at each node, a new approach is proposed for acquisition of load parameter for Now, a network of heterogeneous computational resource. The current methods in use for acquisition of workload are reviewed for comparison and the implemention of prediction model is described in detail. It is concluded from the analysis of experimental results that the prediction method takes less time and has better predictability, and it gives a beneficial support to task scheduling and dynamic load balance.展开更多
文摘In order to solve the problem of significant fluctuation of workload at each node, a new approach is proposed for acquisition of load parameter for Now, a network of heterogeneous computational resource. The current methods in use for acquisition of workload are reviewed for comparison and the implemention of prediction model is described in detail. It is concluded from the analysis of experimental results that the prediction method takes less time and has better predictability, and it gives a beneficial support to task scheduling and dynamic load balance.
