Abstract. In this work we undertake the creation of a framework for testing the degree to which network intrusion detection systems (NIDS) detect and handle evasion attacks. Our prototype system, idsprobe, takes as input a packet trace and from it constructs a configurable set of variant traces that introduce different forms of ambiguities that can lead to evasions. Our test harness then uses these variant traces in either an offline configuration, in which the NIDS under test reads traffic from the traces directly, or a live setup, in which we employ replay technology to feed traffic over a physical network past a NIDS reading directly from a network interface, and to potentially live victim machines. Summary reports of the differences in NIDS output tell the analyst to what degree the NIDS's results vary, reflecting sensitivities to (and possible detections of) different evasions. We demonstrate idsprobe using two popular open-source NIDSs and report on their respective abilities in dealing with evasive traffic.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.