Anomaly detection in VoIP system using neural network and fuzzy logic

Abstract

As Voice over IP (VoIP) system is widely accepted by industries because of its economical benefit and its reach feature than traditional PSTN system. Most dominating session initialization protocol used by VoIP system is SIP, which is vulnerable to various threats like Denial-of-Service (DoS). The Intrusion Detection System (IDS) is use to identify to VoIP threats, but the current IDS system have fail to generate new rule set based on new signature of attack. We have proposed an Adaptive IDS using Artificial Neural Network (ANN) & Fuzzy Approximation technique to detect DoS attacks. Neural Network System is used to learn about new threats while Fuzzy System decides the severity of attack. Our experimental result shows that a combination approach of Neural Network and Fuzzy approximation proposed method have maximize detection rate and minimize false-positive rate. © 2011 Springer-Verlag.