Dynamic Anomaly based System for Real Time Intrusion Detection

As we all know a simple firewall can no longer provide enough scrutiny as in the past. Today's data arrangements in power and business sectors are dispersed whose necessitation requires the use of multiple systems, both proactive and reactive. The premise behind intrusion detection systems is simple: Use a set of agents to filter both network traffic and look for the “signatures� of known network attacks. In this work we have focus on Simple and Hybrid ANN based approach for anomaly detection and how Back Propagation Neural Network (BPNN) cab be used for anomaly detection. One of the novel challenges in the area of IDS detecting unknown or modified attacks because we all know attacks can also have geographically distributed nature. Anomaly based IDS can play a vital role in the field of IDS. We present a Neural Network based technique, which can be used to reduce the false positive in intrusion detection and improve the detection rate in intrusion detection. We will focus on Simple and Hybrid ANN based approach for anomaly detection. This paper will discuss how Back Propagation Neural Network (BPNN) is used for anomaly detection.