A Flexible Load-Balancing Traffic Grooming Algorithm in Service Overlay Network

To overcome the unbalanced flow distribution problem caused by private routing on service overlay network, network service providers usually use load balancing strategy, but traditional methods do not consider making the load balance adapt to the status of the network, neither consider that load balancing may damage performance of some application links. In order to solve the two problems, based on Wardrop equilibrium theory, the paper proposes a flexible load-balancing traffic grooming strategy, which aims to achieve system optimization. According to the principle of stochastic user equilibrium, a network traffic optimization control model is proposed. The path impedance calculation formula is given based on the Logit stochastic add-in term, in which the value of the equilibrium parameter determines the level and scope of load balancing. Finally, we designed the elastic traffic grooming algorithm ElaGo. The simulation results verified the effectiveness of the proposed algorithm.