Q-learning based distributed denial of service detection

<span lang="EN-US">Distributed denial of service (DDoS) attacks the target providers by sending a huge amount traffic to prevent legitimate users from getting service. These become more challenging in software-defined network paradigm, due separation control plane data plane. Centralized software defined networks are vulnerable DDoS that may cause failure all networks. In this work, new approach is proposed based on q-learning enhance detection and reduce false positives negatives. The results work compared with entropy terms number received packets detect attack also continuity for users. Moreover, these indicate system detects flash crowds redirects edge center. A second controller used redirect honeypot server works as mirror server. This guarantees both normal suspected until further analysis done. an increase up 50% throughput other approaches.</span>