A Self-adaptive Predictive Congestion Control Model for Extreme Networks

To combine the design strategies of both preventive control and reactive control, a self-adaptive predictive congestion control model for extreme networks is proposed. First, a model is employed to predict the data traffic. Then the reference trajectory of the traffic is updated, and thereafter transmission flow is adjusted according to the prediction results by optimization of a performance index. The modeling of the system can be performed on-the-fly to accommodate the variation of network traffic and to adapt to the uncertainties of the environment. Simulations are provided to evaluate the system.