IMR: MM-1C: Effective Geolocation of Internet Hosts

点击翻译按钮获取中文摘要

This project tackles the problem of determining the geographic location of Internet hosts from their Internet addresses (the “geolocation problem”). Geolocation is a fundamental problem in computer networking that affects numerous aspects of Internet operation, research, and development. Unfortunately, ample evidence shows that existing geolocation services are prone to errors and inaccuracies, especially when it comes to mapping content provider platforms. Beyond the obvious impact on applications relying on geolocation, these inaccuracies undermine the rigor of any research that utilizes geolocation. This research will develop novel techniques that improve the efficiency and accuracy of geolocation and leverage these techniques to systematically assess the accuracy of influential existing geolocation services. Geolocation commonly involves measuring network delays from a set of “landmarks” (the hosts with known locations) to the geolocation target. By selecting – differently for different targets -- a small number of landmarks from a large pool of candidates, this project attempts to break the dilemma where a large number of landmarks is desired for higher accuracy but leads to scalability limits due to high measurement volumes. Through developing a new geolocation methodology, assessing if its new capabilities could make provably correct geolocation feasible for a substantial fraction of Internet hosts, and evaluating the accuracy of existing geolocation services, this activity will advance our knowledge in the area of Internet host geolocation, a fundamental problem in networking affecting numerous aspects of Internet development.Advancing the state of the art in Internet host geolocation has broad impact through benefiting any applications and research that relies on geolocation. To the extent that it will show that provable geolocation is practical, this project will transform how networking research relying on geolocation is done, by allowing rigorous reasoning about host locations. Datasets containing active measurements, as well as software artifacts, produced in the course of this project will be made publicly available. The project will significantly enhance undergraduate and graduate education at CWRU and facilitate its ongoing efforts in broadening participation of underrepresented sections of the population in computer science.All the data (anonymized when needed) and relevant software artifacts the project produces will be made available through the project website at http://engr.case.edu/rabinovich_michael/geolocation. The content itself will be stored at external public repositories such as github (github.com) and OSF (osf.io), while the project website will provide links to the content along with its description. This will be kept available for at least 5 years after the project completion.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.