Novel Container Cloud Elastic Scaling Strategy based on Kubernetes
Zhigang He · 2020 IEEE 5th Information Technology and Mechatronics Engineering Conference (ITOEC) · 2020
Kubernetes is a cluster management system based on container technology, which is responsible for managing large-scale containers. However, the existing responsive scaling strategy of the Kubernetes platform has a response delay, and the resource indicators cannot meet the web system's requirements for service quality. In order to solve this problem, this paper proposes a new container cloud elastic scaling strategy, which combines a prediction model and a queuing network to automatically predict the arrival rate of requests for web applications in the future and find the optimal number of Kubernetes service replica sets. Finally, in order to prove that our strategy is effective, a Kubernetes cluster is set up in a laboratory environment, and experiments are designed to verify it. The results show that our proposed strategy can achieve predictive scaling and avoid response delay caused by peak requests. At the same time, our elastic scaling strategy can also improve resource utilization to a certain extent.