Scalable Vehicle Location Tracking Using Geohash and Redis

This paper presents scalable algorithms for real-time vehicle location tracking using Redis, an in-memory key-value store. Traditional single-machine solutions, while efficient, do not scale well. This study explored two sharding methods-prefix sharding and vehicle ID sharding-both leveraging geohash to translate two-dimensional spatial data into a one-dimensional space. Prefix sharding splits data by geohash prefixes, while vehicle ID sharding distributes data based on vehicle IDs. Performance tests on rectangle and radius queries reveal that prefix sharding excels with uniformly distributed, small queries, while vehicle ID sharding evens out the load and performs better under non-uniform, large queries. Implementations are benchmarked against Redis' native geo features, showing the strengths and weaknesses of each approach in terms of query response times and throughput.