Performance analysis for a service delivery platform in software defined network

Software Defined Networking (SDN) is an emerging networking paradigm that is expected to play a significant role in future networks. Although exciting progress has been made in research and development of SDN technologies, application of this new networking paradigm in large scale networks with heterogeneous autonomous domains, such as the Internet, is still a challenging issue that calls for a new service delivery platform offering a high-level abstraction for network programmability and inter-domain collaboration. In this paper, we first present a service delivery platform that employs the Network-as-a-Service (NaaS) notion for end-to-end service provisioning in SDN. Then we propose a novel modeling and analysis method for evaluating performance of the service delivery platform. The proposed analysis method addresses the challenges brought in by network abstraction and domain diversity in the service delivery platform by applying network calculus techniques. Both analytical and numerical results obtained in this paper indicate that our modeling and analysis well reflect the advantage of centralized network control enabled by SDN for end-to-end service provisioning; thus achieving tighter performance bound compared to traditional methods for network performance analysis.