SQL vs. NoSQL

One-Liner

The choice between structured, relational databases (SQL) and flexible, non-relational databases (NoSQL).

What It Is

SQL (Structured Query Language): Databases that use a structured schema (tables, rows, columns) and are known for reliability and complex query capabilities (e.g., MySQL, PostgreSQL). They adhere to ACID properties.
NoSQL (Not Only SQL): Databases that do not use a fixed schema and are known for their flexibility, scalability, and performance. They come in various models (key-value, document, column-family, graph) and often favor BASE properties.

Why It Exists

To provide different database models for different needs. SQL is great for transactional, structured data, while NoSQL is built for the scale and flexibility required by modern web applications.

How It Works

SQL: Enforces a predefined schema. Data is normalized to reduce redundancy.
NoSQL: Schema is dynamic. Data is often denormalized for faster reads.

Tradeoffs

SQL

Pros: Strong consistency (ACID), powerful query language, mature ecosystem.
Cons: Difficult to scale horizontally, rigid schema.

NoSQL

Pros: Easy to scale horizontally, flexible schema, high performance for specific access patterns.
Cons: Weaker consistency guarantees (eventual consistency), joins are often not supported.

Failure Modes

SQL: The database becomes a single point of failure and a scaling bottleneck.
NoSQL: Data inconsistency due to eventual consistency, application-level complexity to handle lack of joins.

Interview Traps

Saying one is always better than the other.
Not being able to justify the choice of one over the other for a specific use case.

Real-World Usage

SQL: Financial systems, e-commerce order management.
NoSQL: Social media feeds, IoT data, real-time analytics.

Anti-Patterns

Using NoSQL for a system that requires complex transactions and joins.
Using SQL for a system that has a rapidly evolving data model and needs to scale massively.

ACID vs. BASE
Database Sharding
Read Replicas