Roads Belong in Databases

The popularity of location-based services and the need to perform real-time processing on them has led to an interest in queries on road networks, such as finding shortest paths and finding nearest neighbors. The challenge here is that the efficient execution of operations usually involves the computation of distance along a spatial network instead of “as the crow flies,” which is not simple. This requires the precomputation of the shortest paths and network distance between every pair of points (i.e., vertices) with as little space as possible rather than having to store the n2 shortest paths and distances between all pairs. This problem is related to a ‘holy grail’ problem in databases of how to incorporate road networks into relational databases. A data structure called a road network oracle is introduced that resides in a database and enables the processing of many operations on road networks with just the aid of relational operators. Two implementations of road network oracles are presented.