An experimental exploration of the impact of multi-level packet loss on network intrusion detection

Abstract

In this paper we consider the problem of packet loss as it applies to network intrusion detection. We explore the research question: is the impact of packet loss on network intrusion detection performance sufficiently regular to allow a formula to be developed that will predict the effect? We constructed 2 experimental environments to allow us to measure the impact of packet loss. We graphed the packet loss rate against the alert loss rate. We used nonlinear regression analysis to produce a formula with R squared and adjusted R squared values close enough to 1 for us to answer our research question in the affirmative.