A Virtual Machine Dynamic Adjustment Strategy Based on Load Forecasting.

Uneven assignment of tasks may cause virtual machine (VM) overload or underload in cloud computing environment. No matter overload or underload, the efficiency of cloud resources will be much affected. Especially underload, a lot of resources are not utilized which causes much waste. To solve this problem, a VM dynamic adjustment strategy based on load forecasting is proposed. Through load forecast, the strategy predicts the bottleneck of the key resources that affect the performance of the system. Utilizing the prediction results the resources are dynamically and effectviely adjusted. Extensive experiments show the strategy is correct and efficient. It can much improve the utilization efficiency of resources and lay a foundation for further study of VM adjustment strategy.