A multi-service differentiation traffic management strategy in SDN cloud data center

Abstract

The cloud computing technology makes services more diverse, and users have different needs for multi-service experience. The operator cloud data centers support more services and users, resulting in a rapid growth in traffic. Consequently, the traditional network architectures and technologies can no longer meet these traffic management requirements. To address this problem, this paper proposes a multi-service differentiated (MSD) traffic management strategy for the operator SDN cloud data center. The MSD traffic management rules and model are designed according to the operational requirements in cloud data centers. Fibonacci tree optimization algorithm (FTO) is improved for matching the MSD model to optimize the differentiated traffic management solutions. An operator SDN cloud data center test platform is established to verify the effectiveness of the strategy. The results show that the MSD-FTO strategy can obtain multiple differentiated traffic management schemes through one single optimization and it outperforms ECMP. In addition, it can properly manage the multi-service traffic and greatly improve the network performance and service experience in cloud data centers.