Mapping the Internet: Geolocating Routers by Using Machine Learning

Citations Over Time

Abstract

Knowing the geolocation of a router can help to predict the geolocation of an Internet user, which is important for local advertising, fraud detection, and geo-fencing applications. For example, the geolocation of the last router on the path to a user is a reasonable guess for the user's geolocation. Current methods for geolocating a router are based on parsing a router's name to find geographic hints. Unfortunately, these methods are noisy and often provide no hints. This paper presents results on using machine learning methods to "sharpen" a router's noisy location based on the time delay between one or more routers and a target router or end user IP address. The novelty of this approach is that geolocation of the one or more routers is not required to be known.

Related Papers

• → Bridging router performance and queuing theory(2004)64 cited
• → Bridging router performance and queuing theory(2004)11 cited
• → Overview of router architectures(2004)
• An Approach to Enhance the Security of Router(2006)
• The Attack and Defense of Router(2013)