Towards Web-Scale and Energy-Efficient Hybrid SDNs: Deployment Optimization and Fine-Grained Link State Management

Abstract

As one of the most promising solutions for the next generation Web-scale datacenter networks, hybrid software-defined networks (SDN) have received extensive attention and in-depth research in recent years. Specifically, the issues of deployment optimization and fine-grained link state management in hybrid SDNs are extremely crucial for building Web-scale and energy-efficient SDN-driven datacenters. In this paper, we introduce an energy-efficient deployment optimization scheme for hybrid SDNs, in order to reduce the deployment cost and energy consumption, so that the hybrid networks can approach or even achieve the centralized control ability of pure SDN networks. In addition, we propose fine-grained link state management algorithms with energy consumption sensing ability. With the algorithms, the link state of the hybrid SDNs can be controlled in a fine-grained way, and can be managed by dynamically clearing or temporarily closing redundant links, so as to further reduce link energy consumption. We setup experiments to demonstrate the functionality effectiveness and performance gains of the proposed scheme and algorithms, and the extensive simulation results show that, the proposed solution of this paper outperforms its competitors, producing an arresting increase of energy efficiency. Specifically, under the premise of only 20% of SDN switch deployment budget, the proposed scheme in this paper can reduce the total energy consumption by about 40%.