Cloud Resource Scheduling Algorithm Based on Combination Weight

The default scheduling algorithm in the cloud environment only considers the two performance indicators of CPU and memory, and uses a unified weight to calculate the candidate node score, which cannot meet the needs of different Pod applications. Disk and IO rate are three indicators, and the subjective weight and EW (entropy weight) are calculated by AHP (analytic hierarchy process, analytic hierarchy process), and resources are calculated in real time according to the resource utilization of node performance indicators during Pod application deployment process The objective weight of the indicator. The combination of the two weights is applied to the improved TOPSIS (technique for order preference by similarity to an ideal solution) multiple attribute decision-making method to select suitable candidate nodes. With the increase in the number of deployed Pods, in the case of a large cluster load, the standard deviation of the comprehensive load is significantly improved compared with the default scheduling algorithm of Kubernetes.