I use the JSPRIT with time/distance matrix acquisition, which is a relatively long process no matter what method the matrix acquisition is to be chosen (opensource or commercial e.g. HERE). Long time is being hardly seen when matrix size goes beyond 2000x2000.
ODL Studio uses ROAD_NETWORK built by the graphhopper library based on OSM pbf files. It can uses external matrix files as well.
I wonder how is the extracting (speed and distance data) process being done with this data, which allows ODL Studio to perform so well?