Virtualized Computing and Network Integration in Software-Defined Data Center Networks: An Energy-Efficient Approach

The rapid expansion of integrated computing-networking data centers has led to a corresponding increase in energy consumption, creating significant sustainability challenges. With increasing attention on carbon peaking and carbon neutrality, traditional resource management strategies fall short of the green, low-carbon requirements of modern cloud data centers. To address this challenge, this paper presents an energy efficient data center network framework that consists of Software-Defined Networking (SDN) controller, coordinator, and OpenStack to jointly monitor and manage compute and network resources. The virtual computing resources management model is proved to be a two-dimensional bin-packing problem, which optimizes the dual goals of reducing energy consumption and improving resource utilization. To reduce the computational complexity, we propose an Improved Grey Wolf Optimization-based Virtual Machine Placement (IGWO-VMP) algorithm. For network resource management, we formulate an energy-efficient model based on the 0-1 knapsack problem. By taking advantage of SDN technology, we propose the Dynamic Energy-Efficiency Routing (DEER) algorithm. Numerous experimental results show that the average energy consumption of IGWO-VMP algorithm is reduced by 10.23% compared with the traditional methods. The DEER algorithm has the lowest energy consumption and the highest percentage of energy saving.