Anfis Based Multilayered Botnet Detection with Traffic Reduction Technique
M. Kempanna, R. Jagadeesh Kannan · 2015
2 Abstract: In recent years, Botnets have become the most serious threat for cyber-security. Botnet is a malicious software application which can be controlled by remote system from outside network through command and control channel. Most of the existing behavior based techniques could not able to detect and predict the botnet as they change their structure and pattern. For increasing of the efficiency of botnet detection, the multi-agent systems have been deployed. Intelligent systems with fuzzy and neural-fuzzy techniques improve the botnet presence degree in computer networks. In this paper, Adaptive Neuro Fuzzy Inference System (ANFIS) is used to train the system for future prediction. The multi-layered architecture is combined with ANFIS for the detection of a wide range of existing and new botnets. In addition, to improve the overall system performance, we develop a traffic reduction algorithm to reduce the amount of network traffic required to be inspected by the proposed system. Simulation results show that the proposed system achieves a high detection rate (98.75%) and a low false positive rate. The traffic reduction algorithm reduces an average traffic by 80%.