SSAS Tabular to Fabric: Models at compatibility level 1500 or above can be deployed to Fabric's Analysis Services endpoint with connection updates and validation. For Direct Lake, the model connects to OneLake Delta tables instead of importing data.
SSAS Multidimensional (OLAP) to Fabric: There is no direct import path. The cube architecture must be rebuilt as a Tabular star schema model, with DAX replacing MDX calculations, dimensional hierarchies redesigned, and MDX-consuming reports assessed for compatibility.
Both migration paths require: security redesign, partition assessment, downstream client testing, and parallel validation before decommissioning the existing SSAS environment.
SSAS to Microsoft Fabric migration moves SQL Server Analysis Services semantic models, either Tabular or Multidimensional, to Microsoft's unified analytics platform. Tabular models can be deployed to Fabric's Analysis Services endpoint or converted to Power BI semantic models using Direct Lake. Multidimensional OLAP cubes require a broader redesign, rebuilding dimensions, measures, hierarchies, and MDX logic as a Tabular model using DAX.
The migration scope therefore depends heavily on the source architecture and can include MDX to DAX conversion, hierarchy redesign, SSAS partition migration, role-based security to Fabric RLS, and validation against existing SSAS query results.
The first step in any SSAS to Microsoft Fabric migration is identifying whether the source environment uses Tabular models or Multidimensional cubes. The two architectures have different storage engines, query languages, and migration requirements.
| Dimension | SSAS Tabular | SSAS Multidimensional |
|---|---|---|
| Storage engine | VertiPaq (columnar, in-memory) | MOLAP (multidimensional, cube-based) |
| Query language | DAX (also supports MDX for compatibility) | MDX (native) |
| Data model | Relational tables with relationships | Dimensions, measures, and hierarchies |
| Fabric migration effort | Moderate: deploy or convert | High: full redesign required |
| Direct Fabric path | Deploy to Fabric as endpoint or Power BI semantic model | No direct import; must be rebuilt as Tabular |
For organizations planning a migration, this distinction determines the level of redesign required and the appropriate migration strategy.
Organizations are moving from SSAS to Microsoft Fabric to reduce infrastructure management and consolidate analytics workloads. Traditional SSAS environments require server management, SQL Server Analysis Services installation, service administration, and capacity planning. Fabric provides a managed analytics platform that brings semantic modeling together with Lakehouse, Warehouse, Power BI, and other analytics capabilities.
Direct Lake is another major consideration. Instead of importing data into the VertiPaq engine, a Fabric semantic model can read Delta tables directly from OneLake. This can reduce the processing and refresh requirements associated with traditional Import-mode SSAS Tabular models.
Fabric also connects semantic models with the broader Power BI ecosystem, including Power BI Desktop, paginated reports, Excel PivotTables through XMLA, and other reporting tools. This creates a more unified environment for organizations modernizing their existing SSAS and reporting architecture.
The migration becomes easier to plan when each SSAS component is mapped to its corresponding Fabric implementation.
| SSAS Asset | Microsoft Fabric Equivalent | Migration Consideration |
|---|---|---|
| SSAS Tabular model | Fabric semantic model via XMLA endpoint | Compatibility level 1500+ required for direct deployment |
| SSAS Multidimensional cube | Power BI Tabular semantic model | Full redesign required; no direct import |
| DAX measure (Tabular) | DAX measure in Fabric semantic model | Usually compatible; validate against source output |
| MDX calculated member | DAX measure | Re-derived by business intent rather than syntax translation |
| SSAS dimension and hierarchy | Power BI dimension table and hierarchy | Parent-child hierarchies may require redesign |
| SSAS partition | Fabric semantic model partition or OneLake Delta partition | Direct Lake can reduce partition requirements |
| SSAS database role | Power BI RLS role with DAX filter | Rebuilt per role and validated with representative users |
| SSAS data source | Fabric Lakehouse or Warehouse via OneLake | Connection must be reconfigured |
| SSAS processing schedule | Fabric semantic model refresh schedule | Refresh logic must be rebuilt; Direct Lake can reduce refresh requirements |
This mapping provides a practical starting point for determining which assets can be migrated with limited changes and which require redesign.
MDX to DAX conversion is one of the more demanding parts of SSAS OLAP cube migration because the two languages use different calculation models.
MDX is designed around multidimensional spaces, tuples, sets, and dimension hierarchies. DAX works primarily through tables, relationships, filter context, and evaluation context. Because of these differences, MDX logic is generally re-derived according to its business purpose rather than translated line by line.
| MDX Pattern | DAX Approach |
|---|---|
| Calculated member | DAX measure in the semantic model |
| Named set | Calculated table or DAX filter expression |
| Time intelligence (YTD, MTD) | DAX time-intelligence functions such as DATESYTD, DATESMTD, and CALCULATE |
| PeriodToDate functions | CALCULATE with appropriate date filter expressions |
| Scope assignment | CALCULATE with explicit filter context |
| MDX CROSSJOIN | DAX CROSSJOIN or SUMMARIZE |
| Parent-child hierarchy | DAX PATH/PATHITEM functions or a flattened hierarchy table |
Complex MDX involving subselects, custom member calculations, unary operators, and cell-level security requires expert review. Some calculations do not have a direct DAX equivalent and may require changes to the underlying data model rather than a simple measure conversion.
SSAS Multidimensional to Tabular is the more complex migration path because there is no direct import from a Multidimensional cube into a Tabular model. Instead, you must redesign the cube around a relational star-schema approach.
One major difference is the storage engine. Multidimensional SSAS uses MOLAP, where cube data can be pre-aggregated. Tabular models use VertiPaq and DAX to evaluate calculations against relational tables. As a result, existing performance characteristics cannot simply be assumed to carry over to the new model.
Parent-child hierarchies and custom rollups require particular attention. Multidimensional models can use unary operators and custom aggregation behavior that do not have direct Tabular equivalents. These structures may need to be represented using DAX PATH functions or flattened hierarchy tables.
Security is another important consideration. Multidimensional SSAS can use cell-level security, while Power BI RLS works at the row level. Existing cell-level restrictions therefore need to be redesigned rather than directly copied.
Downstream MDX clients must also be assessed. Excel PivotTables and third-party reporting tools may rely on MDX-specific behavior, including custom members. Test these clients against the migrated Fabric semantic model before cutover.
Many-to-many relationships may also require additional redesign. Tabular implementations commonly use bridging tables and DAX calculation patterns to reproduce relationships that were handled differently in Multidimensional SSAS.
SSAS Tabular to Fabric migration is generally more straightforward than Multidimensional migration, particularly when the source model is already at compatibility level 1500 or above.
You can typically export the model definition using tools such as Tabular Editor or SQL Server Management Studio; update connections to point to Fabric-hosted data or OneLake, and deploy the model to the Fabric workspace through the XMLA endpoint.
Security also needs to be updated. Existing Windows authentication groups can be replaced with Microsoft Entra ID groups, while SSAS roles can be rebuilt as Power BI RLS roles using DAX filters.
Next, select the target storage mode. Organizations can continue using Import mode or move toward Direct Lake when their source data is available as Delta tables in OneLake. After deployment, validate DAX measures and model behavior against the original SSAS environment.
SSAS partitions divide large fact tables into independently processable segments. They are commonly used to refresh recent data without reprocessing an entire historical dataset.
In Fabric, the appropriate partition strategy depends on the semantic model's storage mode. Import-mode models can use semantic-model partitions managed through the XMLA endpoint and tools such as Tabular Editor. You can recreate existing partition definitions and update their data sources for the Fabric environment.
Direct Lake changes this approach because the semantic model reads Delta tables directly from OneLake. In many cases, the storage-layer partitioning of the Delta tables can reduce the need for semantic-model-level partition management.
Organizations moving to Direct Lake should therefore review their existing SSAS partition strategy rather than automatically reproducing every partition in the new environment.
SSAS role-based security must be redesigned when moving to Fabric. The exact implementation depends on how the existing SSAS environment controls access.
| SSAS Security Component | Fabric / Power BI Equivalent |
|---|---|
| Database role | Power BI RLS role with DAX filter expression |
| Dimension data security | RLS DAX filter restricting dimension rows |
| Cell-level security | Redesigned using RLS and Object-Level Security (OLS) where appropriate |
| Windows security group | Microsoft Entra ID security group |
| Data-level user filter | Dynamic RLS using USERPRINCIPALNAME() |
Test each role with representative user accounts before retiring the SSAS environment. This confirms that users continue to see only the data permitted by their existing security model.
Complex MDX calculations involving subselects, custom members, unary operators, and advanced set functions may not have direct DAX equivalents. These calculations require individual analysis and may require data-model changes.
Parent-child hierarchies with custom rollup behavior need to be redesigned for Tabular models. DAX PATH/PATHITEM functions or flattened hierarchy tables can be used depending on the required reporting behavior.
Multidimensional cell-level security cannot be directly replicated through Power BI RLS because RLS operates at the row level. Existing restrictions therefore require a redesigned security model.
Excel PivotTables and third-party tools that query Multidimensional SSAS through MDX may behave differently when connected to a Fabric semantic model. Test these clients as part of the migration.
Tabular models below compatibility level 1500 require an upgrade before direct deployment to Fabric. Validate any compatibility-level change for DAX and model behavior.
Multidimensional MOLAP and Tabular VertiPaq use different approaches to storage and calculation. Workloads that relied on MOLAP pre-aggregation may require DAX optimization or aggregation-table design after migration.
SSAS to Microsoft Fabric Migration by DataTerrain
17 Years Experience 400+ US Clients SSAS Tabular and OLAP MDX to DAX Conversion Direct Lake Migration
DataTerrain is a specialist data engineering and analytics migration company that delivers end-to-end SSAS to Microsoft Fabric migration: Tabular and Multidimensional model assessment, MDX to DAX conversion, Direct Lake configuration, SSAS role-based security to Fabric RLS, partition migration, and parallel validation. Our Automated BI reports conversion service complements SSAS migration for organizations modernizing both analytical models and report libraries simultaneously. Start with a free assessment of your SSAS estate before committing to a full migration.
SSAS to Microsoft Fabric migration is not a single migration pattern. The approach depends primarily on whether the existing environment uses Tabular or Multidimensional SSAS.
Tabular models can often move to Fabric with manageable changes when the compatibility level and data architecture are suitable. Multidimensional OLAP cubes require a more substantial redesign, as Tabular semantic models require MDX to DAX conversion, hierarchy redesign, and security migration that requires particular attention.
A successful migration therefore starts with an accurate assessment of the existing SSAS estate, followed by controlled model migration, security validation, downstream client testing, and parallel query comparison before final cutover.
Contact DataTerrain to discuss your SSAS environment and plan the migration.