Distributed Sensor Objects for Intrusion Detection Systems

Abstract

Effective intrusion detection requires the analysis of enormous volumes of data collected from distributed sensor objects. Traditionally, this analysis has been performed on central servers. The alternative has been to limit data collection to network flow data, with the side effect of reducing intrusion detection effectiveness. This research examined an alternative, namely the incorporation of more intelligent sensor objects. We discuss the infrastructure required to support such a metaphor, the potential tradeoffs, and a novel algorithm for such an intelligent distributed sensor object.