iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Microsoft Fabric is a shared SaaS analytics platform: OneLake provides its tenant-wide logical data lake, workspaces organize items and access, and capacities supply the compute and billing boundary for assigned workspaces. The key design decision is not simply where to put data; it is how to group workloads, permissions, regions, and resource use.
How Microsoft Fabric architecture fits together
A useful way to picture Fabric is as an organization’s tenant containing capacities and workspaces, with data and analytics items inside the workspaces. OneLake spans the tenant as a shared logical data foundation. It is not a separate lake created for every workspace.
- Tenant: The organization’s identity and administration boundary.
- Capacity: A region-bound compute and billing resource identified by a Fabric SKU and capacity units (CUs).
- Workspace: A collaboration, access, and governance container assigned to a capacity.
- Items: The artifacts in a workspace, such as lakehouses, warehouses, notebooks, pipelines, semantic models, and reports.
Fabric brings data integration, engineering, data science, real-time analytics, databases, warehousing, and Power BI into one SaaS environment. Those experiences share platform services and can work with data in OneLake. Microsoft describes the intended pattern as reusing data across workloads rather than making a separate copy for each one; that does not mean every source, operation, or workload is copy-free. Microsoft’s Fabric overview explains the platform’s workload model.
Free tools Windows power users keep installed
One-click scans. No signup required.
What OneLake does—and what it does not mean
OneLake is the logical data lake for a Fabric tenant, built on Azure Data Lake Storage. It provides a common namespace and storage foundation for Fabric workloads. Microsoft documents Delta Parquet and Iceberg formats for tables. Workspace boundaries organize ownership and access within that tenant-wide lake; they do not turn each workspace into an independent tenant lake. See Microsoft’s OneLake overview for the service’s organization and supported formats.
#1 Best Overall
“One copy” is best understood as a reuse goal: Fabric workloads can work against shared data rather than requiring a duplicate for every experience. It is not a guarantee that moving, ingesting, transforming, or serving data never creates another copy. Plan for the actual data flow and storage used by your workloads.
How workspaces and capacities relate
A workspace is where a team organizes Fabric items and manages collaboration and access. It is assigned to a capacity, which supplies compute for the workspaces placed on it. Multiple workspaces can use the same capacity; they do not each require their own capacity. Conversely, an organization can use separate capacities for different workload groups.
Rank #2
This makes workspace and capacity placement both architecture and governance decisions. Sharing a capacity can simplify centralized management, but workloads on it can contend for resources and affect query or job performance. Separating capacities can provide workload isolation or support regional requirements, at the cost of managing more capacity resources. The Azure Architecture Center’s Fabric deployment patterns outlines the tenant, capacity, workspace, and item hierarchy and related deployment trade-offs.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteChoose workspace boundaries for people and governance
Use workspaces to reflect meaningful ownership, collaboration, access, and lifecycle boundaries. Consider who needs to create or administer items, which users need access, and how development and production artifacts should be managed. A workspace boundary is not a substitute for deciding which capacity and region should serve its workloads.
Choose capacity boundaries for compute and placement
Group workspaces on one capacity when shared management and resource use suit their workloads. Consider separate capacities when isolation, regional placement, or distinct operational management matters more. Capacities are region-bound, so placement affects where the assigned compute runs; evaluate regional and data-residency requirements before assigning workspaces.
- Governance and delegation: Decide which teams own workspaces and which administrators manage capacity.
- Workload isolation: Shared capacity means shared compute resources and possible contention; separate capacities can isolate groups.
- Region: Capacity placement is regional, so account for residency and workload location requirements.
- Cost management: Capacity consumption and storage are distinct; evaluate workload usage and available pause or scale choices for the capacity and SKU you select.
- Lifecycle: Align workspace count and boundaries with how teams develop, govern, and operate items.
Lakehouse, warehouse, and Direct Lake are different choices
These terms describe different parts or use patterns of a Fabric solution. A lakehouse and a warehouse are storage and analytics experiences; Direct Lake is a Power BI semantic model storage mode for Delta tables in OneLake.
Rank #4
| Option | What it is | Good fit to evaluate | Important consideration |
|---|---|---|---|
| Lakehouse | File storage and tables in OneLake for lake-oriented engineering. Fabric provisions a SQL analytics endpoint for querying its Delta tables. | Workloads centered on lake-oriented data engineering and files alongside tables. | The SQL analytics endpoint provides a way to query Delta tables; choose it based on the workload and operating pattern rather than assuming it replaces every warehouse need. |
| Warehouse | A relational analytics storage experience in Fabric. | Workloads whose data shape, SQL needs, governance, and operating preferences align with a relational warehouse pattern. | It is not categorically better or worse than a lakehouse; the decision depends on the specific workload. |
| Direct Lake | A Power BI semantic model mode that loads data for interactive analysis from Delta tables in OneLake. | Analysis against OneLake Delta tables when avoiding a full Import-mode copy and refresh workflow is desirable. | It requires Fabric capacity, has SKU-specific limits and guardrails, and still benefits from table tuning. |
Use the workload’s data shape, engineering pattern, SQL requirements, governance, and operational preferences to choose between lakehouse and warehouse. Direct Lake is not a third storage container: it is a semantic model mode that uses OneLake Delta tables. Microsoft’s data storage options and lakehouse-versus-warehouse decision guide provide product-specific comparisons. For Direct Lake requirements and constraints, consult the Direct Lake overview.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchSeparate OneLake storage charges from capacity consumption
Storage billing and Fabric capacity-unit consumption are separate parts of the cost picture. Microsoft documents OneLake storage as pay-as-you-go per GB; that storage charge does not consume Fabric CUs. OneLake transactions do consume capacity units, and capacity usage or storage attribution can depend on where the data resides and which capacity performs the access. Excess consumption can lead to throttling.
Best Value
Do not estimate the total cost from stored gigabytes alone: include capacity use by queries, transactions, and other workloads. Prices depend on the relevant region and SKU, so check current pricing for the specific configuration rather than relying on a general figure. Microsoft’s OneLake consumption guidance and capacity consumption example explain the distinction.
A practical way to plan a Fabric deployment
- Map ownership and access. Identify the teams, administrators, data owners, and audiences that need distinct workspace boundaries.
- Group workloads by resource and region. Decide which workspaces can share compute and which need separation for isolation, regional placement, or management.
- Select the data experience per workload. Choose lakehouse or warehouse according to data shape, engineering and SQL patterns, governance, and operations; use Direct Lake only when its semantic-model role and capacity requirements fit.
- Estimate both cost dimensions. Account separately for OneLake storage and capacity consumption, including transactions and workloads that access the data.
- Validate limits and operational behavior. Check the selected SKU’s Direct Lake limits and guardrails, tune tables where needed, and monitor for capacity pressure that could throttle work.
For a guided hands-on introduction to discovering and connecting to data in OneLake, use Microsoft Learning’s Discover and connect to data in OneLake lab.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
