Text decision tree

Adapted from Chapter 16 of Decision Trees: Frameworks for Technology Choices. Sources reviewed 2026-11-07.

START: Start with the workload and existing estate. Validate the shortlist with a pilot.
Q1. Keeping a working Synapse estate?
  Yes -> Retain and reassess Synapse. Synapse offers SQL, Spark and pipelines. [STOP]
  No -> Otherwise consider the target workload. [Q2]
Q2. Power BI and a managed SaaS data plane?
  Yes -> Evaluate Microsoft Fabric. SaaS analytics, OneLake and Power BI. [STOP]
  No -> Otherwise check the engineering requirements. [Q3]
Q3. Spark engineering or ML is central?
  Yes -> Evaluate Azure Databricks. Spark, ML and Unity Catalog governance. [STOP]
  No -> Otherwise compare the SQL and integration requirements. [Q4]
Q4. Otherwise: compare the remaining needs
  Yes -> Pilot the relevant candidates. Include SQL, integration and skills. [END]
END: Test the shortlisted architecture with representative data, permissions, latency and operating cost.

AI-generated chapter illustration

Conceptual illustration for Choose an Azure data platform

The illustration presents three product gateways and several side panels. It does not reproduce the graph's working-estate stop condition or every pilot gate. Follow the text tree for those conditions.

The illustration was generated in BookWriter. Its visual arrangement is interpretive; use the explicit nodes and routes in the graph and text tree for the decision.

Primary references