NoSQL

NoSQL is an umbrella term for databases that depart from the strict relational, SQL-based, table-and-row model. The category emerged in the late 2000s as web-scale applications needed horizontal scaling and flexible schemas that traditional relational systems struggled to provide. NoSQL is not a single technology but four broadly recognised families.

The four families

• Document. JSON-like documents with flexible schema. Examples: MongoDB, Couchbase, Firestore, DynamoDB (in document mode).
• Key-value. Pure key-to-value lookup, often in memory. Examples: Redis, Memcached, DynamoDB (in key-value mode), Riak.
• Column-family. Sparse, wide tables grouped into column families, optimised for write-heavy and analytical reads. Examples: Cassandra, ScyllaDB, HBase, Bigtable.
• Graph. Nodes and edges with first-class traversal queries. Examples: Neo4j, Amazon Neptune, ArangoDB, TigerGraph, Memgraph.

What NoSQL trades

• Often weaker consistency models (eventual or tunable) in exchange for horizontal scaling and availability.
• Often weaker joins and transactions in exchange for simpler scaling stories.
• Often more flexible schemas at the cost of stricter constraint enforcement.

Modern relational databases have absorbed many NoSQL features (JSON columns, horizontal scaling, multi-region), and many NoSQL systems have added transactions; the boundary is no longer sharp.

🔗

Related Terms MongoDB, PostgreSQL, Redis, Document Database, SQL, ACID

📖

Further Reading The Cost of SQL Habits on MongoDB Infrastructure