ADE Microsoft Fabric

Microsoft Fabric is the next evolution of Azure's analytics platform. It unifies data engineering, data warehousing, real-time analytics, data science, and business intelligence in a single product with a single storage layer. Understanding Fabric is increasingly important for Azure data engineers because many organizations are beginning to adopt or evaluate it.

What is Microsoft Fabric

Microsoft Fabric is a Software-as-a-Service (SaaS) analytics platform that combines capabilities previously spread across multiple separate services — Azure Data Factory, Azure Synapse Analytics, Power BI, Azure Databricks features, and Azure Data Lake Storage — into one unified experience.

Think of the traditional approach as running a business across five separate offices in different buildings. Each building has its own staff, processes, and systems. Coordination between them requires extra effort and time. Microsoft Fabric moves everything into one building — a single platform where all teams and tools share the same data, the same security model, and the same workspace.

OneLake — The Unified Storage Foundation

At the heart of Fabric is OneLake — a single, tenant-wide data lake. There is only one OneLake per organization. Every Fabric workspace and workload reads from and writes to OneLake automatically.

OneLake is built on Azure Data Lake Storage Gen2. All data is stored in Delta Parquet format. This means data stored by Fabric's data engineering tools is immediately readable by Fabric's analytics and BI tools — no copying, no ETL between services.

Shortcuts — Virtual Links to External Data

OneLake Shortcuts create virtual links to data in external locations — Azure Data Lake Storage, Amazon S3, Google Cloud Storage, Azure Blob Storage — without physically copying the data. A shortcut appears as a folder in OneLake but the data stays in the original location. This enables a "single pane of glass" view across multi-cloud and hybrid data without data duplication.

Fabric Experiences — The Tools Inside Fabric

Fabric organizes its capabilities into six "experiences" — specialized interfaces for different roles and workloads. They all share OneLake and the same Fabric workspace.

Data Engineering

The Data Engineering experience provides Spark notebooks, Spark job definitions, and Lakehouses. A Lakehouse is the primary data engineering artifact in Fabric. It combines a data lake (files in OneLake) with a built-in Delta table layer (SQL endpoint) in one object. You write Spark code to transform data in the Lakehouse, and the Delta tables are immediately queryable via SQL without any additional steps.

Data Factory

Fabric includes a full-featured data integration tool identical in capability to Azure Data Factory. Pipelines, Copy Activity, Data Flows, triggers, and connectors all work the same way. This is the migration path for existing ADF users moving to Fabric.

Data Warehouse

Fabric's Data Warehouse is a fully managed SQL-based warehouse with T-SQL support. Unlike Synapse's Dedicated SQL Pool, it does not require capacity planning or DWU provisioning. It scales automatically. Data written to the warehouse is stored as Delta Parquet in OneLake and is accessible from all other Fabric experiences simultaneously.

Real-Time Intelligence

The Real-Time Intelligence experience handles streaming data. Eventstream replaces Azure Event Hubs for capturing streaming events. KQL Database (Kusto Query Language Database) replaces Azure Data Explorer for querying time-series and log data at high speed.

Data Science

The Data Science experience provides notebooks with MLflow integration for training and tracking machine learning models. Models trained here access data directly from OneLake Lakehouses.

Power BI

Power BI is fully integrated into Fabric. Reports and semantic models live in the same Fabric workspace as the data they consume. Direct Lake mode — a new connection type unique to Fabric — lets Power BI query Delta tables in OneLake at Import-mode speed without an actual data import. This eliminates the traditional trade-off between freshness (Direct Query) and speed (Import).

Fabric Licensing — Capacity Units

Fabric uses a capacity-based licensing model. You purchase a Fabric Capacity — measured in Capacity Units (CUs). All workloads across all experiences in Fabric consume from this shared capacity pool.

You can also use Pay-As-You-Go through an Azure subscription. Trial capacities are available for 60 days at no cost — sufficient for learning and proof-of-concept work.

Migrating to Fabric from Existing Azure Services

Organizations using Azure Synapse Analytics and Azure Data Factory can migrate to Fabric incrementally:

• ADF pipelines migrate directly to Fabric Data Factory — same interface, same connectors
• Synapse Spark notebooks migrate to Fabric Notebooks — same PySpark code runs without modification
• ADLS Gen2 data lakes connect to OneLake via Shortcuts — no data migration required initially
• Power BI workspaces move into Fabric workspaces — reports and datasets remain unchanged

When to Use Fabric vs Individual Azure Services

Key Points

• Microsoft Fabric unifies data engineering, warehousing, real-time analytics, data science, and Power BI in one SaaS platform
• OneLake is the single shared storage layer — all Fabric workloads read from the same data without copying
• Use Shortcuts to connect OneLake to external data in AWS S3 or Azure Data Lake without moving the data
• Direct Lake mode in Power BI delivers Import-mode speed on Delta tables in OneLake — no data import required
• The Lakehouse combines file storage and Delta SQL tables in one object — the primary data engineering artifact in Fabric
• Existing ADF and Synapse workloads migrate to Fabric with minimal code changes