Approximation of Router Geolocation and Route Visualization

Visualizing traceroutes and the underlying internet infrastructure has been a topic of research for the past decades. Being able to accurately and dependably perform IP geolocation and computations of paths through the Autonomous Systems making up the internet is one part of that. As manually interpreting the output generated by traceroutes is difficult and time-intensive, many tools providing easy-to-use visualizations were released previously. Since accuracy, coverage, or scalability is often an issue with these tools, GeoTrace was developed to try and overcome these shortcomings. This work presents the extension of the existing application. Enhancing the approximated Autonomous System networks used for the traceroute visualizations and adding multiple new sources and techniques in combination with an algorithm comparing and assessing the quality of each source on demand to perform more accurate IP geolocation with great scalability. Additionally, to give users a feel for the accuracy, several additions were made to the user interface that present information about the confidence in the used approximations and sources. The applicability of the final prototype was evaluated using a group of experts working in the domain and provided positive feedback and ideas for further development. Additionally, a technical evaluation of my DNS geolocation algorithm gave a good indication of the impact of using a multi-source system instead of trusting a single source.