Do you see me now? Sparsity in passive observations of address liveness

摘要: Accurate information about address and block usage in the Internet has many applications planning allocation, topology studies, simulations. Prior studies used active probing, sometimes augmented with passive observation, to study macroscopic phenomena, such as overall of IPv4 space. This paper instead completeness sources: how well they can observe microscopic phenomena within a given network. We define sparsity limitation monitor see target, we quantify effects interest, temporal, coverage sparsity. To sparsity, introduce inverted analysis, novel approach that uses complete observations few end networks (three campus our case) infer what these would be seen by millions virtual monitors near their traffic's destinations. Unsurprisingly, find popular content more targets visibility is strongly influenced bipartite traffic between clients servers. are first show implications for liveness from observations. heavy-tailed, only 0.5% seeing than 10% targets' addresses, most affected interest over temporal Visibility also bipartite. Monitors different class target (e.g., server observing client target) outperform same 82–99% cases datasets. Finally, adding probing greatly improves both but not critical blocks. Our findings valuable understand limitations existing measurement develop methods maximize future studies.