Dynamic Resource Allocation Algorithm for Container-based Service Computing

Cloud computing and virtualization technologies play important roles in modern service-oriented computing paradigm. More conventional services are being migrated to virtualized computing environments to achieve flexible deployment and high availability. We introduce a schedule algorithm based on fuzzy inference system (FIS), for global container resource allocation by evaluating nodes' statuses using FIS. We present the approaches to build containerized test environment and validates the effectiveness of the resource allocation policies by running sample use cases. Experiment results show that the presented infrastructure and schema derive optimal resource configurations and significantly improves the performance of the cluster.