A reinforcement learning approach to virtual machines auto-configuration
LI Wen-cha · Electronic Design Engineering · 2014
Virtual machine technology enables multiple virtual machines shared resources on the same physical host. In response to the application requirements change, or changes in supply of resources, distribution of resources on the virtual machine should be able to be reconfigured dynamically. In this article, we propose an algorithm based on reinforcement learning to automate the VM configuration process,Its name is SRLAC(Standard Reinforcement Learning Auto-Configuration). SRLAC emphasizes the algorithm based-model to solve the system management applications to ensure stability and adaptability. Through a virtual machine-based cloud testbed CloudSim implementation of a representative server load experiments prove the validity of the results. This method can find an optimal(or near optimal) allocation strategy in small-scale systems, and it perform a good stability and adaptability.