Skip to main content

Prepare your Fabric landing zone for migration

In the context of Microsoft Fabric, preparing a landing zone for migration primarily involves getting your on-premises or legacy data estate ready for integration or transition. This applies to a wide range of enterprise systems including SQL Server, Oracle, SAP, MySQL, PostgreSQL, file-based sources like Excel or CSVs, APIs, and third-party SaaS applications commonly hosted in shared drives, internal data servers, or enterprise systems.

Objectives​

  • Enable data ingestion from on-premises and legacy systems to Fabric Lakehouses or Warehouses.
  • Integrate and/or modernize diverse data sources via Fabric Pipelines and Dataflows.
  • Leverage Azure-hosted services (e.g., Azure SQL MI, Azure Files, or Azure Blob Storage) as intermediate steps if a full migration is not yet feasible.

Key Preparation Tasks​

1. Evaluate your on-premises sources​

  • Inventory database systems including SQL Server, Oracle, MySQL, PostgreSQL, and SAP systems (e.g., SAP HANA, SAP BW).
  • Identify other data sources such as REST APIs, flat files, data lakes, and third-party SaaS applications.
  • Document authentication mechanisms (Windows Auth, SQL Auth, OAuth, Kerberos, API keys, etc.).
  • Check connectivity constraints (firewalls, VPN, proxy servers).

📘 See Azure Data Migration Guide

2. Enable secure hybrid connectivity​

Use one of the following options to connect on-premises to Microsoft Fabric:

  • On-premises Data Gateway (Enterprise mode) – Recommended for stable, secure, and managed access to on-premises sources.
  • Azure ExpressRoute or VPN – For high-throughput or low-latency needs when integrating Fabric with Azure-hosted services like Azure SQL DB.

📘 See Install and configure an on-premises data gateway

📘 See Hybrid networking guidance in Azure

3. Plan your identity integration​

  • Use Microsoft Entra ID (formerly Azure AD) for identity federation and service principal authentication.
  • Ensure that your Fabric workspace users are Entra-ID integrated for future role-based access control (RBAC).
  • Optionally, extend your AD DS domain to Azure via Azure IaaS if required for legacy apps during migration.

📘 See Identity and access best practices

4. Prepare Fabric workspaces and lakehouses​

  • Define Fabric capacities and assign workspaces to appropriate capacity SKUs.
  • Create Lakehouses or Warehouses in each target workspace.
  • Ensure appropriate roles (Contributor, Admin) are assigned using Fabric RBAC.

📘 See Microsoft Fabric workspace planning

5. Enable ingestion mechanisms​

Depending on your architecture and readiness:

  • Use Fabric Pipelines to pull data from SQL Server, Oracle, MySQL, PostgreSQL, SAP, REST APIs, Flat Files, or Azure Blob into Lakehouse.
  • Leverage Dataflows Gen2 for low-code ingestion and transformation pipelines.
  • Use Notebook-based ingestion via PySpark or SQL for heavy or custom ETL/ELT logic.
  • Utilize generic connectors available via Fabric Pipelines such as REST, SFTP, OData, and others.
  • Consider batch vs. streaming ingestion based on data freshness and volume requirements.

📘 See Ingest data into Microsoft Fabric

6. Optional: Migrate databases to Azure first​

If latency or security concerns prevent direct on-prem → Fabric ingestion:

  • Migrate databases to Azure SQL Managed Instance or Azure SQL DB.
  • Ingest from these Azure services via Direct Lake or Fabric Pipelines.

📘 See Migrate SQL Server to Azure SQL Managed Instance

📘 See Data migration strategy and tools

7. Prepare data for AI and advanced analytics readiness​

Structured data can serve as a powerful foundation for AI use cases when properly modeled, cleansed, and contextualized.

  • Assess data quality and completeness, especially for entities that drive prediction or classification.
  • Identify opportunities to unify related datasets through master data management or entity resolution.
  • Evaluate feature extraction and enrichment options in Fabric (e.g., semantic models, calculated columns, or Python-based feature engineering).
  • Ensure semantic clarity through documentation and consistent naming conventions in your Lakehouse or Warehouse.
  • Use tagging and lineage tracking to clarify data provenance and transformation steps.

📘 See Introduction to AI-ready data preparation in Fabric

📘 See AI and data governance in Fabric

Summary​

You don't need to fully modernize your on-premises or legacy environments before using Fabric. Fabric lets you ingest and model data directly from a broad range of sources with minimal friction using secure connectivity, governed workspaces, managed ingestion pipelines, and advanced data preparation capabilities. Incorporating AI readiness into your data preparation ensures your Fabric landing zone is well-positioned for advanced analytics and AI-driven insights.

Contributors