Abstract: At present the data transmission and communications commonly use both wired and wireless networks. Frequent detection systems are available more for wired networks than for wireless networks. The detection system developed for wired networks cannot be deployed in wireless networks due to the difference between the two types of networks. The wireless detection system is different from the conventional wired detection models. The data transmission of the wired network is a standard physical routing. However, the data stream routing of wireless network are based on radio signals with a variety of problems of evolution. So in both the wired and wireless networks of the internet scenario, it is necessary to deploy detection system efficiently from attackers. The proposed research offered in this thesis is based on anomaly detection of the statistical traffic which is carried out on the normal and abnormal anomaly traces in the packet header and traffic volume detection. The attackers packet header data are acknowledged with a port number, option field parameters and IP address. The anomalies detection is carried out at a regular interval to monitor the traffic analyzed through statistical variance. The change detection detects the statistical variance of the traffic volume. The proposed anomaly traffic detection systems provide a better detection system compared with the existing anomaly detection system. The results obtained from the proposed system are compared with the existing attack detection systems with the propagation delay metric. It shows a reduction of nearly 10% and an improvement of 13% average throughput.
K. Saravanan and R. Asokan, 2016. An Investigation on Neuro-Fuzzy Based Alert Clustering for Statistical Anomaly of Attack Detection. Asian Journal of Information Technology, 15: 136-141.