A Lagrangian Approach to Storage and Access Methods for Spatio - Temporal Network Datasets : A Summary of Results Draft - April 29 , 2010

Spatio-temporal networks are becoming popular with the rise of traffic information being incorporated into street mapping utilities such as Google Maps. This additional information adds potentially large time series of traffic information to each road segment. Naive implementations of secondary storage of spatio-temporal networks may not be sufficient, as current techniques commonly store entire time series with an edge, or store networks a snapshot at a time. Queries such as route evaluation will require a different node and different time-step for each segment of a route, resulting in excessive disk I/O. We introduce a Lagrangian-Connectivity Framework for storing spatio-temporal networks based on each node’s spatio-temporal connectivity.