Multi-objective Container Consolidation in Cloud Data Centers

In recent years, container-based clouds are becoming increasingly popular for their lightweight nature. Existing works on container consolidation mainly focus on reducing the energy consumption of cloud data centers. However, reducing energy consumption often results in container migrations which have big impact on the performance (i.e. availability) of applications in the containers. In this paper, we consider container consolidation as one multi-objective optimization problem with the objectives of minimizing the total energy consumption and minimizing the total number of container migrations within the certain period of time and present an NSGA-II based algorithm to find solutions for the container consolidation problem. Our experimental evaluation based on the real-world workload demonstrates that our proposed approach can lead to further energy saving and significant reduction of container migrations at the same time compared with some existing approaches.