An efficient botnet detection system in large scenario networks using soft computing techniques

Abstract

Cyber security is the major issue in inter and networking due to handling of large stream of multimedia data between nodes in network. The security threat is one of the major issues in present networks which are considered as the most important in handling of multimedia data in networks. Bots are the kind of security attack which attacks the individual or group of nodes in the networks and these affected nodes becomes bots. Botnet is a group of bots which can control by a bot master in network. These bots are instructed to faulty the network activities. It captures the high sensitive information such as bank account details and personal information. At present, botnets are detected using Intrusion detection system (IDS) which effectively monitors the network traffic regularly in entities and enterprises. Bots generates dummy or unwanted information and these fake information are passed to all the nodes in the network, which degrades the network efficiency. After having detailed literature review it is understood that identifying the botnets in a network is a challenge. Hence in this research work it is proposed, Adaptive Neuro Fuzzy Inference System (ANFIS) classifier is used to detect the botnets in the network system. The proposed system extracts the features such as Trust features, Neighborhood features and Entropy features from each node of the network. The features from normal node and botnet affected node are significantly different with each other. These feature set are trained and the trained pattern is used for classification of nodes behaviour into normal or botnet affected node. The performance of this proposed system is evaluated using False Positive Rate and False Negative Rate. newline newline newline