GUI: A Geolocation Method for Unreachable IP.

The article focuses on GUI, a novel IP geolocation method designed to accurately locate unreachable IP targets—those devices or intermediate routers that do not respond to detection messages—at the city and street levels without user assistance. GUI improves city-level geolocation by extracting and comparing metropolitan area network (MAN) topology with target detection paths, then refines location estimates using landmarks within the same /24 subnet or connected via common routes, employing a minimum covering circle (MCC) approach to infer precise positions. Experimental evaluation across 15 cities in China and the USA demonstrates that GUI achieves a city-level geolocation success rate of 99.41% for reachable IPs and 89.37% for unreachable IPs, with a median urban internal geolocation error of 4.99 km, outperforming existing methods such as PoP-Based Geolocation (PBG), Street-Level Geolocation (SLG), Ranking Nodes-Based Geolocation (RNBG), and commercial databases like IPIP and MaxMind. The study also notes that GUI’s accuracy decreases in smaller cities due to fewer landmarks and highlights future work to enhance performance in regions with limited landmark data.