Fabric Foundations and OneLake
Understand the unified Fabric platform, OneLake storage, and how the core workloads fit together
Microsoft Fabric at a Glance
Microsoft Fabric is an all-in-one analytics platform that brings data ingestion, engineering, warehousing, real-time analytics, data science and Power BI into one SaaS environment. Its shared storage foundation is OneLake.
Core Fabric Workloads - Data Factory: ingest, transform and orchestrate data - Data Engineering: build lakehouses and Spark-based transformation workflows - Data Warehouse: create relational analytical stores with transactional T-SQL - Real-Time Intelligence: ingest, store, query and act on streaming events - Data Science: explore data and build machine-learning solutions - Power BI: model, visualise and deliver business insights
Roles Across Data Teams - Data engineer: ingests and transforms data using pipelines, notebooks and lakehouses - Data scientist: uses notebooks, Spark and Python for exploration and machine learning - Data analyst: builds semantic models and reports, often using Direct Lake - Analytics engineer: curates reliable business-ready data and reusable semantic definitions
OneLake: One Logical Data Lake
OneLake is the unified storage layer for Fabric. Data is stored in open formats such as Delta-Parquet so multiple engines, including Spark, SQL and Power BI, can work with the same data without unnecessary duplication.
Key OneLake Concepts - Shortcuts reference data in locations such as Amazon S3 or Azure Data Lake without copying it - Workspaces organise items for collaboration, access control and lifecycle management - The OneLake catalog supports discovery, metadata and governance across the data estate
Memory Tip - Think of OneLake as the shared foundation, while Fabric workloads are different engines and experiences operating on that foundation
Workspaces and Platform Enablement
Fabric must be enabled by an administrator. Workspaces then act as collaboration containers and can provide lineage, Git integration and managed compute. Access can be granted broadly through workspace roles or narrowly through item permissions.
Common Confusion - A workspace is an organisational and security boundary - OneLake is the storage foundation - A lakehouse, warehouse, eventhouse or semantic model is an item inside a workspace