Geospatial Indexing

Your ASP.NET Core API stores each driver's live location as two plain columns, Latitude and Longitude, in Azure SQL. To match a rider with the nearest driver, it loads every driver currently online, works out the distance to each one in C#, and picks the smallest.

At 300 online drivers, that's under a second. But at 4,000 drivers during rush hour, the same query and loop takes over 4 seconds, and it just keeps getting slower as more drivers come online.

A search for what's nearby ends up checking everyone, not just who's close. That's exactly what geospatial indexing fixes.

Geospatial indexing lets the database do that work instead: change the Latitude/Longitude columns to a single geography column in Azure SQL, and create a spatial index on it, so Azure SQL organizes locations by where they are instead of scanning every row in order. Query using STDistance() with a radius filter, so the database returns only drivers within range, already sorted by distance.

Use: Azure SQL Spatial Index (STDistance)

Search only what's nearby, not everyone who exists.

A phone book sorted alphabetically by name is great for finding one person, but useless for finding every restaurant within a mile of you. You would have to read every single entry to check.

A map sorted by neighborhood solves that instantly. You only ever look at the neighborhoods near you, never the whole book.

A map beats a phone book for finding what's nearby, every time.