Generating representative traffic for intrusion detection system benchmarking

In this paper, a modeling and simulation framework is proposed for generating data for training and testing intrusion detection systems. The framework can develop models of Web usage from Web server logs in a data driven fashion and the actual traffic is generated by employing the Web browser installed on the host. Additionally, we employed an intrusion detection system as a traffic analyzer to validate the synthetic data that the framework generated and compared it against the standard intrusion detection system benchmark data, namely KDD 99 datasets.