Introduction
SQL Server Analysis Services is one of those technologies that refuses to fade away. Every year, someone claims SSAS is dead, then I walk into a client environment and see massive tabular models powering mission critical dashboards. Even though the cloud ecosystem has exploded with options like Fabric, Power BI datasets, BigQuery BI Engine, and even DuckDB for embedded analytics, SSAS still finds a place in enterprise architectures. This raises the simple question many teams ask today:
Is SSAS still the right choice, or is it time to move on?
This post breaks down what SSAS looks like today, what new features still matter, why it remains behind in some areas, and what the realistic alternatives are depending on your organization.
Where SSAS Stands Today
SSAS comes in two flavors.
• Multidimensional (classic cubes)
• Tabular (xVelocity based column store engine)
Most new implementations are tabular only. Multidimensional is in long term slow maintenance mode.
Recent and Notable SSAS Features
Even though the product is not aggressively evolving, a few improvements have landed over the last couple of years.
1. Improved Compatibility With Power BI Desktop
You can now author Power BI semantic models offline and deploy them back into SSAS Tabular structures with higher compatibility.
Ref: Microsoft Docs, SSAS Tabular compatibility levels
2. Enhanced Dynamic Management Views (DMVs)
Microsoft continues to add DMV coverage for better monitoring, model health checks, and memory usage breakdowns. This directly helps DBAs who manage large tabular models.
Ref: DMV reference for SSAS
3. Better Integration With Azure AS Migration Tools
Although Azure Analysis Services is being replaced by Power BI Premium datasets, the migration tooling improvements have benefited SSAS as well.
Ref: AAS to PBI migration guidance
4. Ongoing Engine Optimizations
Compression improvements and slight boosts in processing performance continue to trickle down.
Nothing huge but meaningful for high cardinality tables.
Ref: SSAS release notes
5. Extended TMSL and JSON Scripting Support
The Tabular Model Scripting Language continues to expand, making CI or automation a lot smoother for teams using DevOps pipelines.
Ref: TMSL documentation
So is SSAS actively developed?
Not really. SSAS is in a sustained engineering state. That means bug fixes, security updates, and minor improvements. Almost all innovation has moved to Power BI Semantic Models which are built on the same engine.
Why SSAS Is Still Used Everywhere
Even with slow updates, SSAS shows up because of real world operational reasons.
1. Predictable On Prem Licensing
Some enterprises cannot or will not move to the cloud. SSAS offers predictable, server based licensing that avoids per user or per capacity cloud pricing.
Ref: SQL Server licensing guide
2. Deep Enterprise Security and Role Based Access
SSAS row level security is still cleaner and more predictable than the same feature in some cloud BI platforms.
Ref: SSAS RLS documentation
3. Large Semantic Models With Long History
Companies have ten years of investment and visualizations built on SSAS. Migrating a 500 gig model with custom calc groups, partitions, automation, and ML scoring is not a weekend job.
4. Tabular Models Still Perform Exceptionally Well
When properly partitioned, processed incrementally, and optimized, SSAS Tabular remains one of the fastest semantic engines available.
Ref: VertiPaq engine papers
5. IT Governance Requirements
Some industries require the BI layer to be controlled by IT rather than a self service cloud experience.
6. Integration With Legacy Reporting and Line of Business Apps
A lot of SSRS and legacy tools still hook directly into SSAS cubes.
Where SSAS Falls Behind
This is the part most architects care about.
1. Cloud Native Innovation Has Passed It
The most advanced features in the ecosystem are now only available in Power BI Premium and Fabric Semantic Models. Examples:
• Direct Lake
• Real time semantic models
• Auto aggregations
• AI powered modeling
• One semantic layer across the lake
Ref: Microsoft Fabric semantic layer docs
2. Limited Future Roadmap
Microsoft has made it clear. Future semantic engine innovation ships in Power BI and Fabric. On prem SSAS only receives stability improvements.
3. Less Flexible Scaling
SSAS scales vertically.
Power BI and Fabric scale horizontally and elastically.
For workloads with unpredictable spikes, SSAS becomes harder to manage.
4. DevOps and Automation Still Require More Work
Tools like Tabular Editor, ALM Toolkit, and Azure DevOps help, but it still requires more custom scripting compared to modern cloud native semantic platforms.
5. Harder To Integrate With Streaming and Real Time
Most modern BI requirements involve event based ingestion, streaming metrics, and low latency dashboards. SSAS was never designed for real time analytics.
Alternatives To SSAS Today
1. Power BI Premium Semantic Models
This is the true successor to SSAS Tabular.
For most organizations, this is the recommended path forward.
2. Microsoft Fabric + Direct Lake
Direct Lake is the future of high performance analytics in the Microsoft ecosystem.
You get:
• Lakehouse scale
• Tabular engine performance
• No refresh needed
• Unified security
• Enterprise level governance
Ref: Direct Lake announcement
3. Azure Analysis Services
Still supported but not recommended for new projects. Microsoft wants customers to move to Fabric or Power BI.
4. DuckDB for Embedded Analytics
Smaller teams love DuckDB because it runs locally, in process, and hits column store level speeds.
Great for:
• Local analytics
• Application embedded BI
Not a full semantic model though.
5. BigQuery BI Engine or Looker Semantic Layer
For Google Cloud environments, these provide similar capabilities to tabular models.
6. Databricks and Unity Catalog Semantic Layer
If your team is deep into Delta Lake, the Databricks semantic layer becomes attractive.
Is SSAS Still the Best Option?
Here is the honest answer.
SSAS is still the best option if:
• You need on prem only.
• You have existing investment in SSRS, SSIS, and SSAS ecosystems.
• You already have thousands of measures, calc groups, partitions, scripts, and automation that would cost millions to rewrite.
• Your security model is tied tightly to SSAS roles.
• Your performance is already great.
SSAS is not the best option if:
• You want future proof innovation.
• You want to remove refresh cycles.
• You are moving to a lakehouse architecture.
• You need real time analytics.
• You want unified governance like Purview + Fabric.
• You want easier DevOps and lifecycle management.
Workshop Section: Hands On With SSAS Today
This workshop walks through a real world scenario for teams still supporting SSAS but planning a future transition.
Step 1. Inspect the Model Using DMVs
SELECT *
FROM $SYSTEM.DISCOVER_STORAGE_TABLE_COLUMNS
WHERE TABLE_NAME = 'Sales';This helps you see column cardinality, compression, and whether your tables are well optimized.
Step 2. Check Measure Dependencies
Run this to see which measures depend on each other.
SELECT *
FROM $SYSTEM.TMSCHEMA_MEASURES;Step 3. Evaluate Processing Strategy
Review partitioning using the following DMV.
SELECT *
FROM $SYSTEM.TMSCHEMA_PARTITIONS;Step 4. Review Memory Usage
SELECT *
FROM $SYSTEM.DISCOVER_MEMORY_USAGE;Step 5. Test Migration Using ALM Toolkit
• Compare your SSAS model to a blank Power BI semantic model.
• Validate breaking changes.
• Review transformation logic.
Step 6. Benchmark Direct Lake vs SSAS
If you want a real migration decision, run the same queries on Direct Lake.
In almost every case, complex DAX will run several times faster.
Summary
SSAS is a stable, predictable, and still widely deployed enterprise semantic engine. It is not dead and it is not going away tomorrow, but the innovation has moved to Power BI and Microsoft Fabric. The choice between SSAS and modern alternatives depends entirely on your company’s governance, architecture, budget, and migration appetite.
If you want the full power of the Microsoft ecosystem, the future is the Fabric semantic layer and Direct Lake. If your environment is on prem or heavily invested in SSAS, staying where you are is perfectly fine as long as you understand the trade offs.
Final Thoughts
SSAS deserves respect. It powered enterprise BI for more than 20 years and still runs some of the biggest semantic models in the world. But you need to align your strategy with Microsoft’s roadmap. That means planning a slow, controlled move toward Power BI semantic models or Microsoft Fabric over time.
If you want, I can turn this into:
• A full SQLYard formatted article
• A multi slide deck
• A downloadable PDF or Word doc
• A hands on workshop with screenshots
Just tell me what version you want next.
References
- SSAS Tabular compatibility levels
https://learn.microsoft.com/sql/analysis-services/tabular-models/compatibility-level-for-tabular-models - SSAS Dynamic Management Views
https://learn.microsoft.com/sql/analysis-services/instances/analysis-services-dynamic-management-views - Azure AS to Power BI migration
https://learn.microsoft.com/power-bi/enterprise/service-azure-analysis-services-migration - SSAS Release Notes
https://learn.microsoft.com/sql/analysis-services/release-notes - TMSL Documentation
https://learn.microsoft.com/sql/analysis-services/tmsl/tabular-model-scripting-language-tmsl-reference - SQL Server Licensing Guide
https://learn.microsoft.com/sql/sql-server/licensing-guide - SSAS Row Level Security
https://learn.microsoft.com/sql/analysis-services/tabular-models/role-membership-and-permissions - VertiPaq Engine
https://learn.microsoft.com/power-bi/guidance/whitepaper-vertipaq - Microsoft Fabric Semantic Layer
https://learn.microsoft.com/fabric/get-started/semantic-models - Direct Lake
https://learn.microsoft.com/fabric/data-engineering/direct-lake-overview
Discover more from SQLYARD
Subscribe to get the latest posts sent to your email.
