Network Topology Boundary Routing IP Identification for IP Geolocation

Fuxiang Yuan, Fenlin Liu, Rui Xu, Yan Liu, Xiangyang Luo

Abstract: The only few existing methods for network topology boundary routing IP identification are often based on a single network characteristic, with poor applicability, low accuracy, and difficult to meet the IP geolocation requirements. This paper proposes a network topology boundary routing IP identification algorithm for IP geolocation. The proposed algorithm is no longer limited to a specific delay distribution used in D-Based, but identifies the boundary routing IP according to the difference between each single-hop delay in the path and the city delay threshold; more different from traditional methods, when the delay difference characteristic is not obvious, the proposed algorithm combines the differences of routing IP hostnames in different cities to further identify the boundary routing IP and the proportion of identifiable paths is increased. Experiments are carried out using the probing results of millions of landmarks in 10 target cities in China and the US. Results show that compared with D-Based and P-Based, the ratio of identifiable paths is increased by about 103.0% and 69.5%; the number of nodes in the target city is about 1.7 and 1.9 times, and the accuracy of the node location is improved by about 48.3% and 16.9%; using the topology obtained by the proposed algorithm, the success rate of three classic street-level IP geolocation methods such as SLG, LENCR, and Geo-RMP is increased by 75.8%, 61.2%, and 71.7%.