Analytical Subnetwork and IP Alias Resolution for Network Tomography Using Path Traces

Abstract

Mapping of link-level network topologies requires the processing of collected raw traces, including the resolution of alias IPs of router interfaces and underlying subnetworks. There have been several probing techniques for IP alias resolution along with an analytical resolution approach which relies on the IP address assignment practices. The analysis of IP address assignments can also reveal the underlying subnetworks, i.e., the link-level connectivity among the routers. In this paper, we present a comprehensive Analytical Subnetwork and Alias IP Resolution (ASIAR) method that relies on the analytical analysis of path traces to infer underlying subnetworks and router interfaces. While ASIAR increases the efficiency of the analytical resolution method, it integrates additional sanity checks and performs parameter adjustment based on the topology sampling characteristics to improve the resolution accuracy. We explore how different network sampling issues affect the analytical resolution, and analyze the accuracy of the ASIAR on synthetic and genuine networks. Compared with the state-of-the-art analytical resolution method, ASIAR is able to increase both precision and recall by fine-tuning the parameters of sanity checks used for analytical subnetwork and IP alias resolution.