A Kubernetes Autoscaler Based on Pod Replicas Prediction
Tengfei Hu, Yannian Wang
Abstract
Kubernetes, as the de facto standard for scheduling and scheduling containers, has been widely used in the industry. However, the existing scaling strategy of Kubernetes is only the response scaling of thresholds. To improve the efficiency of resource scaling, an active Kubernetes auto scaling device based on Pod replica prediction is proposed. The experiment shows that the autoscaler proposed in this paper has the advantage of faster response speed.
Topics & Concepts
ScalingReplicaComputer scienceScheduling (production processes)Point of deliveryDe factoDistributed computingScalabilityParallel computingMathematical optimizationMathematicsDatabaseAgronomyGeometryPolitical scienceVisual artsLawArtBiologyCloud Computing and Resource ManagementDistributed and Parallel Computing SystemsParallel Computing and Optimization Techniques