Design of optimal acceptance sampling plan for network intrusion detection

Abstract

Network Intrusion Detection Systems (NIDS) protects networks connected to the internet from malicious attacks by monitoring network flows predominantly at fragment level in network layer. Inspecting every fragment of a network flow is computationally prohibitive. The Acceptance Sampling for Network Intrusion Detection (ASNID) method avoids hundred percent inspections of fragments to detect anomalous flows. This study proposes a model to determine optimal acceptance sample size. Further, this study also proposes a model for estimating the cost of computational effort.