A Learning-Based Approach for IP Geolocation

Abstract

Abstract. The ability to pinpoint the geographic location of IP hosts is compelling for applications such as on-line advertising and network attack diagnosis. While prior methods can accurately identify the location of hosts in some regions of the Internet, they produce erroneous results when the delay or topology measurement on which they are based is limited. The hypothesis of our work is that the accuracy of IP geolocation can be improved through the creation of a flexible analytic framework that accommodates different types of geolocation information. In this paper, we describe a new framework for IP geolocation that reduces to a machine-learning classification problem. Our methodology considers a set of lightweight measurements from a set of known monitors to a target, and then classifies the location of that target based on the most probable geographic region given probability densities learned from a training set. For this study, we employ a Naive Bayes framework that has low computational complexity and enables additional environmental information to be easily added to enhance the classification process. To demonstrate the feasibility and accuracy of our approach, we test IP geolocation on over 16,000 routers given ping measurements from 78 monitors with known geographic placement. Our results show that the simple application of our method improves geolocation accuracy for over 96% of the nodes identified in our data set, with on average accuracy 70 miles closer to the true geographic location versus prior constraintbased geolocation. These results highlight the promise of our method and indicate how future expansion of the classifier can lead to further improvements in geolocation accuracy.

On the prevalence and characteristics of MPLS deployments in the open internet

“…Prior work on using Naive Bayes methodologies on Internet measurements have been explored in the context of IP geolocation in [18]. Our approach differs here through the use of path-based measurements (as opposed to only end-to-end measurements in the prior work) and application (MPLS identification vs. geolocation).…”

Understanding geolocation accuracy using network geometry

“…Constraint-Based [5] 42 42 Europe 95 95 N. Amer. Naive Bayes [6] 225 13499 N. Amer. Street Level [7] ∼76000 {72,88} N. Amer.…”

“…Considerable prior work has been done on the subject of measurement-based geolocation [2], [3], [4], [5], [6], [7]. We are motivated by the fact that each of these cited works has used a different geolocation data set to test their methodology, and therefore an 'apples-to-apples' comparison across different studies is not possible.…”

A Location Privacy Aware Network Planning Algorithm for Micromobility Protocols

“…In the current Internet architecture (which also plays as the basis for all-IP mobile and wireless communication systems), an IP address not only identifies a node (or an interface) on the network but also serves as the essential element for routing packets through the Internet topology. Accordingly, when an IP packet is sent from one Internet node to another, both sender and receiver entities reveal their topological location (i.e., their IP addresses) in the network, which can then easily be translated to a quite accurate estimation of the peers' current geographical location (Lakhina, Byers, Crovella, & Matta, 2003), (Freedman, Vutukuru, Feamster, & Balakrishnan, 2005), (Gueye, Ziviani, Crovella, & Fdida, 2006), (Baden, 2008) (Eriksson, Barford, Sommersy, & Nowak, 2010), and thus making third parties able to track mobiles' real-life movements or posing other threats to users . In order to protect location privacy in next generation networks, several ideas, schemes and protocols have already been proposed in the literature.…”