Empirical Evaluation of the Internet Analysis System for Application in the Field of Anomaly Detection

Abstract

Anomaly detection in computer networks is an actively researched topic in the field of intrusion detection. The Internet Analysis System (IAS) is a software framework which provides passive probes and centralized backend services to collect purely statistical network data in distributed computer networks. This paper presents an empirical evaluation of the IAS data format for detecting anomalies, caused by attack traffic. This process involved the generation of labeled evaluation data based on the 1999 DARPA Intrusion Detection Evaluation data sets and two different supervised machine learning approaches for the assessment. The results of this evaluation conclude, that the IAS is not a convenient data source for advanced anomaly detection in the scope of our research.