Microsoft Fabric Overview – Complete Guide for DP-900
Microsoft Fabric Overview
Why Is Microsoft Fabric Important?
Microsoft Fabric represents a major evolution in how organizations approach analytics on Azure. It is important because it unifies multiple data services — data engineering, data integration, data warehousing, real-time analytics, data science, and business intelligence — into a single, integrated platform. Before Fabric, organizations had to stitch together separate Azure services (Azure Data Factory, Azure Synapse Analytics, Azure Databricks, Power BI, etc.) to build end-to-end analytics solutions. Fabric simplifies this by providing a cohesive experience built on top of a shared foundation.
For the DP-900 exam, understanding Microsoft Fabric is critical because Microsoft is positioning it as the future of analytics workloads on Azure. Exam questions increasingly reference Fabric as the unified analytics solution, and understanding its components, architecture, and purpose helps you answer a wide range of questions about modern data analytics.
What Is Microsoft Fabric?
Microsoft Fabric is an end-to-end, unified analytics platform delivered as a Software-as-a-Service (SaaS) offering. It brings together the following key workloads under a single platform:
1. Data Engineering – Build and manage data pipelines and transformations using Apache Spark.
2. Data Integration – Use Data Factory experiences within Fabric to create ETL/ELT pipelines for moving and transforming data.
3. Data Warehousing – Create and manage relational data warehouses with T-SQL support.
4. Real-Time Intelligence – Ingest and analyze streaming and real-time data using KQL (Kusto Query Language) databases and eventstreams.
5. Data Science – Build, train, and deploy machine learning models using notebooks and experiments.
6. Business Intelligence (Power BI) – Create reports, dashboards, and interactive visualizations directly within the Fabric environment.
All of these workloads are built on top of a shared foundation that includes:
- OneLake – A single, unified data lake for the entire organization. OneLake is built on Azure Data Lake Storage Gen2 and serves as the centralized storage layer. All Fabric workloads read from and write to OneLake, eliminating data silos.
- Unified governance and security – Managed through Microsoft Purview integration, providing consistent data governance across all workloads.
- Unified capacity model – A single billing and compute model that can be shared across all Fabric workloads.
How Does Microsoft Fabric Work?
OneLake – The Foundation:
At the heart of Fabric is OneLake, which functions like a "OneDrive for data." Every Fabric tenant gets a single OneLake instance. Data stored in OneLake is organized into lakehouses and warehouses. OneLake uses the open Delta Parquet format, ensuring data is stored in an open, interoperable format. This means data is accessible across all Fabric workloads without needing to copy or move it.
Lakehouses:
A lakehouse in Fabric combines the flexibility of a data lake with the analytical power of a data warehouse. It allows you to store structured, semi-structured, and unstructured data and query it using both Spark and SQL. The lakehouse automatically creates a SQL analytics endpoint so you can query data using T-SQL without having to move it to a separate warehouse.
Workspaces and Experiences:
Fabric uses workspaces as containers for organizing items (datasets, reports, pipelines, notebooks, etc.). Users switch between different experiences (Data Engineering, Data Warehouse, Power BI, etc.) within the same Fabric portal. This unified experience means a data engineer, a data scientist, and a business analyst can all collaborate on the same data without leaving the platform.
Capacities:
Fabric uses a capacity-based licensing model. Organizations purchase Fabric capacities (measured in Capacity Units or CUs), and these capacities are shared across all workloads. This simplifies cost management compared to provisioning separate resources for each service.
Data Integration with Pipelines:
Fabric includes Data Factory pipelines (similar to Azure Data Factory) that allow you to orchestrate data movement and transformation. You can use dataflows (Power Query-based) and pipelines (activity-based orchestration) to ingest data from hundreds of sources into OneLake.
Key Concepts to Remember:
- SaaS platform – Fabric is fully managed; you don't need to provision individual Azure resources.
- OneLake – Single, unified storage layer; no data duplication needed across workloads.
- Open formats – Data is stored in Delta Parquet format, promoting interoperability.
- Unified governance – Integrated with Microsoft Purview for data cataloging, lineage, and sensitivity labels.
- Multi-persona – Supports data engineers, data scientists, data analysts, and business users in one platform.
- Shortcuts – OneLake supports shortcuts, which are pointers to data in external storage (like Azure Data Lake Storage Gen2 or Amazon S3) without copying the data.
How Microsoft Fabric Fits Into the Analytics Workload on Azure:
In the context of DP-900, analytics workloads involve ingesting, processing, storing, and visualizing data. Microsoft Fabric consolidates all of these steps:
- Ingestion: Data Factory pipelines and dataflows bring data into OneLake.
- Storage: OneLake stores all data in a unified lake.
- Processing: Spark notebooks, T-SQL queries, KQL queries, and ML models process and transform data.
- Visualization: Power BI reports and dashboards present insights to business users.
This end-to-end integration is what makes Fabric a comprehensive solution for modern analytics.
Exam Tips: Answering Questions on Microsoft Fabric Overview
1. Know that Fabric is SaaS, not PaaS or IaaS. If a question asks about a fully managed, unified analytics platform, Fabric is the answer. It differs from Azure Synapse Analytics (which is more PaaS-oriented) in that Fabric requires no individual resource provisioning.
2. OneLake is the single most important concept. Expect questions about what OneLake is, how it eliminates data silos, and how it stores data in open Delta Parquet format. Remember: one tenant = one OneLake.
3. Understand the distinction between a lakehouse and a warehouse in Fabric. A lakehouse combines lake flexibility with warehouse querying (supports both Spark and SQL). A warehouse in Fabric is a traditional relational data warehouse that uses T-SQL exclusively.
4. Remember the unified experience. Questions may describe a scenario where multiple personas (data engineer, analyst, data scientist) need to collaborate. Fabric is the answer because it provides a single platform for all these roles.
5. Shortcuts are a key feature. If a question mentions accessing external data without copying it into OneLake, the answer involves OneLake shortcuts.
6. Capacity-based licensing. If asked about how Fabric is billed or how compute resources are shared, remember the capacity model. One capacity can serve multiple workloads.
7. Do not confuse Fabric with individual Azure services. Fabric includes capabilities similar to Azure Data Factory, Azure Synapse, and Power BI, but it is a unified platform, not just a bundle of these services.
8. Governance integration. If a question mentions data governance, lineage, or sensitivity labels in the context of Fabric, the answer typically involves Microsoft Purview integration.
9. Real-Time Intelligence. If a question describes streaming or real-time data analysis within Fabric, look for answers mentioning Eventstreams and KQL databases.
10. Elimination of data silos. A recurring theme in exam questions is the problem of data silos. Fabric's OneLake and unified architecture are designed specifically to solve this problem. If a question asks how to reduce data duplication or unify disparate data sources, Fabric/OneLake is likely the correct answer.
11. Watch for distractor answers. Questions might list Azure Synapse Analytics, Azure Data Lake Storage, or standalone Power BI as options. While these are valid Azure services, if the question emphasizes a unified, end-to-end, SaaS analytics platform, the answer is Microsoft Fabric.
12. Use process of elimination. If you see a question about analytics and the options include both "Microsoft Fabric" and individual services, consider whether the scenario requires integration across multiple workloads. If yes, Fabric is almost always the best answer.
By understanding these core concepts — OneLake as the foundation, the unified multi-workload experience, SaaS delivery, open data formats, and capacity-based compute — you will be well-prepared to answer any DP-900 exam question related to Microsoft Fabric.
