Current country borders storage holds for each edge an entry consisting of a pair of country IDs at its ends, plus the type of border.
Proposed solution
The storage size could be greatly reduced (by a factor of at least three) by switching from an edge-based to a node-based encoding such that only a single country ID is stored for each node. Border crossings could be then resolved on the fly in the edge filter by comparing the country entries at its ends. In case of a border its openness could be queried from a precomputed map.
Problem description
Current country borders storage holds for each edge an entry consisting of a pair of country IDs at its ends, plus the type of border.
Proposed solution
The storage size could be greatly reduced (by a factor of at least three) by switching from an edge-based to a node-based encoding such that only a single country ID is stored for each node. Border crossings could be then resolved on the fly in the edge filter by comparing the country entries at its ends. In case of a border its openness could be queried from a precomputed map.