NoSQL

NoSQL refers to a family of database management systems that diverge from the traditional relational model. They are designed to scale horizontally across commodity hardware, handle large volumes of structured, semi-structured, or unstructured data, and often use flexible or dynamic schemas. NoSQL is a broad term that encompasses several data models beyond tables, rows, and fixed SQL schemas.

The main NoSQL data models are key-value stores, document stores, column-family stores, and graph databases. Key-value stores offer simple lookups by a key. Document stores manage semi-structured documents, typically in JSON-like formats, with indexing on document fields. Column-family stores organize data into families of columns, enabling wide rows and sparse schemas. Graph databases represent entities as nodes and relationships as edges, optimized for traversals.

Historically, NoSQL emerged in the mid-2000s as a response to the constraints of traditional RDBMS on scalable, web-scale workloads and big data. Many NoSQL systems emphasize availability and partition tolerance over strict consistency, aligning with BASE rather than ACID principles. The CAP theorem remains a guiding consideration when choosing a system, with different databases offering varying consistency guarantees and transaction support.

Typical use cases include real-time analytics, content management, session stores, and IoT data ingestion. NoSQL databases are popular for high write throughput and flexible schemas, and they commonly run on commodity hardware or in cloud environments. Notable examples include MongoDB (document), Cassandra and HBase (column-family), Redis (key-value), Neo4j (graph), Couchbase, and DynamoDB.