Determining the geographic location of nodes enables a variety of useful network applications. Such applications include finding servers that are closest to clients, customizing content based on the location of website visitors, locating faulty nodes during network diagnosis, and tracking down persons of interest. Past efforts to solve the geolocalization problem have either relied on databases of questionable accuracy and integrity or on limited sources of information about node location. As a result, their coverage and accuracy have suffered. This thesis presents Alidade, a distributed framework for performing accurate and scalable geolocalization with worldwide coverage. Alidade provides a principled approach to geolocalization based on geometric constraint satisfaction. It takes advantage of diverse sources of ground truth, including landmarks whose positions are approximately known. The framework partitions constraints extracted from the network into disjoint sets that can be independently and concurrently evaluated. This enables a MapReduce style implementation that scales well with the number of landmarks and targets.