Design of an Efficient Entropy-based DDoS Attacks Detection Scheme in Software Defined Networking

In recent times, the occurrence of DDoS attacks has been increasing, thereby it boosts up the vulnerability of SDN’s security. These attacks have a damaging impact on the controller resources of SDN, impeding its capacity to effectively process incoming data. The network operations can be disrupted and legitimate users may face difficulties which access essential network services. It becomes crucial to prioritize the protection of the SDN controller, especially against highly developed exploits that specifically exploit the distinctive attributes of SDN. An entropy-based DDoS attack detector is proposed that uses the KNN classification algorithm. The KNN model’s accuracy is found to be 98.0% which indicates that it can correctly classify 98% of the samples. The model’s false alarm rate was found to be 0.13% which again indicates that it is capable of detecting DDoS attacks with high accuracy and reducing the number of false positives.