A delay deviation tolerance IP geolocation method with error estimation

Abstract

AbstractIP geolocation is an important basis of location-based network services, while error estimation is an important basis for judging the reliability of results. Most of the existing IP geolocation algorithms cannot estimate the geolocation error. A few can achieve error estimation through high-precision delay measurement, but their performance is also affected by the common delay inflation in the actual network. A new IP target location estimation method is proposed in this manuscript to achieve geolocation with reliable error estimation of IP targets in actual network. Firstly, after the landmark set divided into training set and verification set for path detection, the metropolitan area network (MAN) topology is extracted through train path set. Secondly, the governed landmarks are searched level by level through the MAN, and the minimum covering circles are calculated through the geographical distribution of the landmarks to infer the routers’ area center. Then, geolocation errors are counted after simulated geolocation through the verification path set, and the minimum mean square error radius of the error mean and the minimum covering circle radius is calculated as the router area radius. Finally, the path to the IP target is measured and compared with the MAN to get the location estimation result. The experimental results based on 12 cities in China show that compared with the existing typical algorithms, the proposed method not only improves the error estimation accuracy, but also has finer geolocation granularity and lower median error.