A Novel Resource Scheduling Approach in Container Based Clouds

Cloud computing is a new paradigm to deliver computing resources to customers in a pay-as-you-go model. In this paper, the recently emerged container-based virtualization technology is adopted for building the infrastructure of a cloud data center. Cloud providers are concerned with the resource usage in a multi-type resource sharing environment while cloud customers desire higher quality of services. According to this, we propose a resource scheduling approach for the container virtualized cloud environments to reduce response time of customers' jobs and improve providers' resource utilization rate. The stable matching theory is applied to generate an optimal mapping from containers to physical servers. Simulations are implemented to evaluate our resource scheduling approach. The results show that our approach achieves a better performance for customers and maximal profits for cloud providers by improving the resource utilization.