Data-Driven Assessment of IP Geolocation Accuracy Using Hybrid Active–Passive Measurement Techniques

Abstract

Proper IP geolocation is essential to the areas of cybersecurity, content delivery, and enforcing compliance regionally. Nevertheless, traditional geolocation databases are usually affected by irregularities including dynamism of IP allocation, asymmetry of the route, and noise on measurements. This paper suggests a hybrid approach to be used in measuring and enhancing the accuracy of IP geolocation through a combination of active delay-based probing and passive data correlation. The study measures the distribution of spatial error and patterns of cross-continent deviation by using a dataset of more than 1.2 million traceroute and latency observations made in the vantage points throughout the world, paired with the data of Border Gateway Protocol (BGP) and Autonomous System (AS) metadata. The experimental findings reveal that the hybrid model proposed will generate an accuracy rate 28 higher than the conventional database searches. The analysis also shows regional biases where routing aggregation and lack of infrastructure transparency in the developing regions make the error rates higher. An artificial intelligent correction model was created, which uses location deviation prediction based on spatial and time scales and with a mean error of less than 25km. The results underscore the opportunity of hybrid geolocation systems to facilitate Internet measurement systems and offer a platform of real-time optimization of accuracy in both business and academic setting.