Choosing the Right Database

#131 - Data Essentials in Software Architecture - Pramod Sadalage

Topic 4 of 8

Published May 1, 2023 · Clip 24:24–30:30 (6:07)

Pramod lays out a thought process for choosing the right database. First, productivity: human time costs more than technology, so e.g. document databases fit JSON-heavy NodeJS work. Second, analyze use cases: scale needs (trillions of rows, distribution/sharding) and read/write patterns, noting disk space is no longer a constraint but read and write efficiency are. Third, consider what you want out of the database (analytical aggregations, graph traversal, etc.). He explains that classically there are five broad types (relational, key-value, document, wide-column, graph) and recommends setting up a scoring rubric (familiarity, cloud alignment, open source, distribution) and running POCs to score candidates like Neo4J, Neptune, DGraph, and TigerGraph.