Predictive Container Auto-Scaling for Cloud-Native Applications
Hanqing Zhao, Hyunwoo Lim, Muhammad Hanif, Choonhwa Lee · 2019
In the past decade, cloud computing has become an essential technology in many areas such as Internet of Things, artificial intelligence, and social media. In the cloud-computing environment, the auto-scaling capability of services is important to optimize cloud operating costs and Quality of Service. Therefore, there is a need for auto-scaling technology that is able to dynamically adjust resource allocation to cloud services based on incoming workload. In this paper, we present a predictive auto-scaler for Kubernetes clusters to improve the efficiency of container auto-scaling. Being based on a predictive algorithm, our auto-scaling scheme simplifies the architecture of existing auto-scaling system for more efficient service offerings. In addition, we present experimental evaluation results of our proposed scheme.