How does graphhopper load worldwide routing network. i want to load worldwide network and how do i do that. do i do that in a single machine. or partition the data. if i partition the data how does the routing network know to go from point a to point b in different partitioned data?
There are some instructions in the GitHub repo on how to deploy it at world scale here
They even mention machine specs required. From what I read, I don’t think partition is an option but I’m relatively new to this
You should be able to simply load planet.osm.pbf into the graphhopper configuration (see last reply) in order to allow it to process planet-wide requests. You do however need a very large amount of RAM (about 128-196 gigabytes) to do this properly.
(Partitioning could be done to decrease RAM usage if you wanted to do it, but you’d have to partition by, e.g., continent or landmass, because you wouldn’t be able to handle routes that go between partitions. https://download.geofabrik.de/ has “extracts” of parts of planet.osm.pbf that you can process if you wanted to do that. You’re probably better off loading planet.osm.pbf directly.)