MAGNA: modeling and generating network attacks

Abstract

We present a system that generates synthetic attack traffic from state-based models of the behavior of real network attacks. The execution of these attacks can be carefully controlled to produce realistic training data that could prove useful in the evaluation and development of intrusion detection systems This tool can also be used to test host systems and networks for known vulnerabilities by launching controlled attacks and observing their impact on the target.