Resource Based Automatic Calibration System (RBACS) Using Kubernetes Framework

Abstract

Kubernetes, a container orchestrator for cloud-deployed applications, allows the application provider to scale automatically to match the fluctuating intensity of processing demand. Container cluster technology is used to encapsu-late, isolate, and deploy applications, addressing the issue of low system reliabil-ity due to interlocking failures. Cloud-based platforms usually entail users define application resource supplies for eco container virtualization. There is a constant problem of over-service in data centers for cloud service providers. Higher operating costs and incompetent resource utilization can occur in a waste of resources. Kubernetes revolutionized the orchestration of the container in the cloud-native age. It can adaptively manage resources and schedule containers, which provide real-time status of the cluster at runtime without the user’s contribution. Kuber-netes clusters face unpredictable traffic, and the cluster performs manual expansion configuration by the controller. Due to operational delays, the system will become unstable, and the service will be unavailable. This work proposed an RBACS that vigorously amended the distribution of containers operating in the entire Kubernetes cluster. RBACS allocation pattern is analyzed with the Kuber-netes VPA. To estimate the overall cost of RBACS, we use several scientific benchmarks comparing the accomplishment of container to remote node migra-tion and on-site relocation. The experiments ran on the simulations to show the method’s effectiveness yielded high precision in the real-time deployment of resources in eco containers. Compared to the default baseline, Kubernetes results in much fewer dropped requests with only slightly more supplied resources.