In the evolving landscape of databases, the lines distinguishing one type from another have blurred as features converge. How do you pick the right on

Choosing a backend database: SQL vs Document vs Columnar

submited by

Style Pass

2024-03-31 16:00:05

In the evolving landscape of databases, the lines distinguishing one type from another have blurred as features converge. How do you pick the right one for your needs? Here's a venn diagram showing which features are shared and which are unique.

Schema Flexibility: While the allure of schemalessness might steer you towards NoSQL, keep in mind that modern SQL databases offer JSON column support.

Write-Heavy Workloads: If your primary operation is writing and you require scalability, a columnar DB might be your best bet.

Whether or not one uses a NoSQL database, many of the principles of NoSQL/document databases are helpful when designing an application. For example, ask yourself, how could I design a system where I wouldn’t need to use joins (see Data Denormalization)

One can vertically scale instances quite a lot these days, for example AWS supports up to 128 vCPU individual instances, but that is still only so much.