To determine the location of a computer on the Internet without resorting to outside information or databases would greatly increase the security abilities of the US Air Force and the Department of Defense. The geographic location of a computer node has been demonstrated on an autonomous system (AS) network, or a network with one system administration focal point. The work shows that a similar technique will work on networks comprised of a multiple AS network. A time-to-location algorithm can successfully resolve a geographic location of a computer node using only latency information from known sites and mathematically calculating the Euclidean distance to those sites from an unknown location on a single AS network. The time-to-location algorithm on a multiple AS network successfully resolves a geographic location 71.4% of the time. Packets are subject to arbitrary delays in the network; and inconsistencies in latency measurements are discovered when attempting to use a time-to location algorithm on a multiple AS network. To improve accuracy in a multiple AS network, a time-to-location algorithm needs to calculate the link bandwidth when attempting to geographically locate a computer node on a multiple AS network. | Geographic Location Of A Computer Node Examining A Time-to-location Algorithm And Multiple Autonomous System Networks by Duane C Sorgaard