A Multidimensional Cost Model for Data Models Denormalization

Traditional joins in the relational world are problematic in NoSQL. The fact that data is distributed over the network creates considerable network communications that did not exist in a centralized system.

The aim of these exercises is to show how important it is to model data collections correctly in a distributed context. Schema denormalization is particularly addressed to avoid having to perform these joins.

This post was preliminary posted in 2017, updated since then. This new version was updated in late 2024 to include :

• A presentation for JSON & JSON Schema
• Joins, Denormalization & The Multidimensional Cost Model
  • A complete multi-dimensional cost-model (time, financial, environmental)
  • Update the complexity of the approach
  • Links to ModelDrivenGuide & Fact-DM

Lien vers les Vidéos de correction