A comparative study for accuracy of anomaly detection methods of adaptive flow counting in SDN

The monitoring of network traffic helped in ensuring the integrity of the data inside the network. Software defined networking (SDN) provided a platform for network administrators to easily apply monitoring policies on the networks. SDN´s centralized control and programmability features aids in efficient monitoring of network traffic in distributed environments. Various efficient anomaly detection techniques using adaptive monitoring have already been proposed by many researchers. Different results for anomaly detection were obtained on applying different updated algorithms using adaptive monitoring. In this paper, results of anomaly detection method using adaptive flow counting are compared upon using 1) Reduced Complexity algorithm for dynamic rule update, 2) Dynamic Threshold Range Calculation algorithm for anomaly detection and 3) Improvised Performance algorithm for anomaly detection. It also found the best scenario for accurate detection of anomalies while considering the performance of the network.