Kubernetes-based Workload Allocation Optimizer for Minimizing Power Consumption of Computing System with Neural Network

Edge computing has been attracting attention due to the spread of the Internet of Things. For edge computing, containerized applications are deployed on multiple machines, and Kubernetes is an essential platform for container orchestration. In this paper, we introduce a Kubernetes based power consumption centric workload allocation optimizer (WAO), including scheduler and load balancer. By using WAO built with power consumption and response time models for actual edge computing system, 9.9% power consumption was reduced compared to original Kubernetes load balancer. This result indicates that the WAO developed in this study exhibits promising potential for task allocation modules as a micro service platform.