On The Power And Limitations Of Detecting Network Filtering Via Passive Observation

Network operators often apply policy-based traffic filtering at the egress of edge networks. These policies can be detected by performing active measurements; however, doing so involves instrumenting every network one wishes to study. We investigate a methodology for detecting policy-based service-level traffic filtering from passive observation of traffic markers within darknets. Such markers represent traffic we expect to arrive and, therefore, whose absence is suggestive of network filtering. We study the approach with data from five large darknets over the course of one week. While we show the approach has utility to expose filtering in some cases, there are also limits to the methodology.