Google’s distributed NoSQL database for large-scale structured data is called Bigtable.
Google developed Bigtable beginning in 2004 to store enormous datasets across many machines while maintaining high performance and availability. The system was described in the influential 2006 paper “Bigtable: A Distributed Storage System for Structured Data,” written by Google researchers including Jeffrey Dean and Sanjay Ghemawat.
Bigtable is a wide-column database rather than a traditional relational database. It organizes information by row keys, column families, and timestamps, allowing sparse data and multiple historical versions of individual values. Tables are divided into smaller units called tablets, which can be distributed across servers and rebalanced as the dataset grows.
Bigtable influenced many later storage systems and has supported Google products such as web indexing, Google Maps, Google Analytics, and Gmail. It is often confused with Spanner, Google’s globally distributed relational database, or LevelDB, a separate embedded key-value store. Bigtable is specifically associated with large-scale structured data and NoSQL storage.