Databases
NoSQL Database
A database that trades rigid tables and schemas for flexible data models (documents, key-value, graph), often to scale writes horizontally.
NoSQL covers the databases that stepped away from the relational table model. Document stores like MongoDB keep flexible JSON-like records. Key-value stores like Redis map keys to values at blistering speed. Wide-column and graph databases handle their own specialized shapes. The shared thread is trading strict schemas and cross-table joins for flexibility or horizontal scale.
The pitch is scaling writes across many nodes and storing data whose shape changes often, without schema migrations. That flexibility is genuinely useful for certain workloads: high-volume event logging, session storage, caching, real-time feeds. Many providers offer managed versions, and Redis-compatible caches show up on nearly every VPS host's menu.
The common mistake is reaching for NoSQL by default because it sounds modern. Most applications have relational data and are better served by SQL. Choose NoSQL when your access patterns and scale genuinely demand it, not as a starting assumption.
Go deeper
- Managed Database vs Running Your Own: The Real MathA managed PostgreSQL costs 2-4x the raw server underneath. Whether that premium is worth it depends on exactly one question: what your untested backup is worth.