Distributed Denial of Service(DDoS) attacks are inevitable. The existing defensive mechanisms are relatively outdated. In this paper, we present a passive mechanism to reduce the impact of an attack on the network. We designed and implemented a robust feedback architecture, to maintain the stability of the network despite attacks. During an attack, the controller of the feedback architecture detects the irregularities in the response time and the necessary changes are made to the configuration to maintain the network in steady state. In this approach first, we model the network using black-box system identification technique. Second, we validate the model with test data by conducting various experiments such as varying the network topology. Last, we test the model with the feedback architecture built in our lab environment. Results show that the feedback architecture provides an average model fit accuracy with positive results.