A Traffic Scheduling Scheme for Data Center Networks Based on SDN

As the scale of data center networks (DCNs) continues to expand, software-defined network (SDN), as an innovative network architecture, gradually applies to data centers. SDN realizes flexible control of network traffic by separating control from forwarding. The traditional traffic scheduling algorithm in the network cannot meet the requirements of DCN service, and cannot fully utilize the redundant bandwidth in DCN topology. Therefore, in this paper, for the specific topology and traffic characteristics of DCNs, we propose a traffic scheduling scheme based on SDN. The scheme initially aggregates the elephant flows with the same source and destination, and then exploits the improved fruit fly optimization algorithm to uniformly schedule the aggregated elephant flows. The simulation results show that compared with the classical traffic scheduling algorithm, the improved fruit fly optimized scheduling algorithm (IFOSA) effectively improves the network performance, especially in terms of load balancing, average throughput, and link bandwidth utilization.