Meadow Ride asks why nearby-driver results sometimes point at an offline driver or assign one driver twice. Its fictional fleet has 24,000 active drivers, each reporting every 4 seconds. Deliver a location pipeline and a checked recommendation covering freshness, reordered points, geographic boundaries and exclusive matching.
Meadow Ride needs nearby-driver results for its fictional city. The app receives frequent location updates, but mobile networks can delay or reorder them. A driver who appeared nearby thirty seconds ago may be somewhere else now. Treat a coordinate as a timestamped observation with a version, not a permanent fact.
A geographic cell index narrows the candidates. It also creates a boundary case: the nearest driver may be just across the edge of the rider's cell. Query neighboring cells and calculate distance on the candidates. Matching still needs an authoritative claim so two riders can't both receive the same driver. In this case, build the location pipeline and choose a recommendation that preserves freshness without storing every historical point in the serving cache.