Infrastructure Capability Aggregation and Exposure in Cloud Environments

Abstract

In today’s large-scale distributed cloud environments, there are usually multiple layers involved when performing orchestration. It is important to decide what capability is visible at a specific physical location in the different layers of abstraction. Considering the diversity and heterogeneity of the infrastructure, it is critical to decide how data centers/clusters aggregate and expose their capabilities to the user to be accessible from the orchestration environment. In this paper, we propose a capability aggregation and exposure method. The proposed method relies on the detailed information of the underlying resource which can be extracted using OS built-in tools. It then aggregates and exposes this information using different algorithms. Our proposed method has many advantages such as applicability and flexibility. We evaluated the performance of the different algorithms and demonstrated the one that has the best performance.