Virtual machine anomaly detection strategy based on cloud platform operating environment perception

摘要

In this paper, a variety of anomaly detection strategies are provided with the application ofthe cloud platform center in view of the processor resources allocation according to the peakload demand, provision of the single anomaly detection strategy, and the problem of the suddenincrease of the abnormal rate due to the change of the resource demand. The gray waveformdetection algorithm is applied to detect the anomaly arriving at the virtual machine in the futuretime period, and the abnormal utility function of the virtual machine is given to balance theresource requirements and anomaly detection priority, and all the virtual machines are dynamicallyconfigured for each virtual machine with the maximization of the anomaly utility value ofthe virtual machine as the objective. Through the global load balancing between the same virtualmachines and the redistribution of the virtual machines for multiple times, the detection quantityon the virtual machine anomalywith relatively huge increase of anomaly is further increased.Finally, the anomaly detection algorithm based on the gray waveform detection in the cloud platformcenter ADGWT is given. Simulation experiment results show that the proposed algorithmcan effectively improve the abnormal rate of the processors in the cloud platform, and it haspractical significance to improve the completion rate of the user requests.