Power BI Premium provides the richest set of capabilities for both self-service and enterprise BI solutions. Azure Analysis Services is also an option for enterprise BI solutions. So if you’re looking at deploying an enterprise BI solution, which one is for you?
First – some definitions. Power BI Premium is a hosted service (SaaS) that allows you to host datasets and share Power BI reports and dashboards with anyone in your organization without requiring recipients to be individually licensed. Power BI Premium provides dedicated hardware and capacity hosted by Microsoft for reliable and consistent performance. This dedicated capacity can be scaled up or down according to your business needs. In addition to datasets, Power BI Premium includes several other workloads including dataflows, paginated reports, and AI. Analysis Services is the engine powering the datasets workload.
Analysis Services tabular models can be created and deployed in Azure Analysis Services. With Azure Analysis Services, almost all tabular models can be moved into Azure with few, if any, changes. There are multiple deployment options and service tiers within each option that you can tailor to meet your requirements. In less than a minute, you can configure the service tier you require in the Azure Portal. Note that multi-dimensional models are not yet supported in Azure Analysis Services.
Power BI Premium is the focus for enterprise BI and the primary target for future investments. In time, Power BI Premium will provide a superset of the capabilities when compared to Azure Analysis Services.
With a few exceptions, Power BI Premium provides a superset of the capabilities available in Azure Analysis Services. In time, these exceptions will be eliminated making Power BI Premium a clearly superior choice when considering capabilities alone.
The table below lists where the significant differences exist between the two offerings:
|Power BI Premium||Azure Analysis Services|
|Unlimited Power BI content viewing||Yes||No|
|Multi-model memory management||Yes||No|
|Automated incremental refresh||Yes||No|
|Large datasets||Yes (in preview)||Yes|
|Third party application support||Yes*||Yes|
|Bring Your Own Key (for encryption)||Yes (in preview)||No|
|Scale out||Not yet||Yes|
|Metadata Translations||Not yet||Yes|
|Object Level Security||Not yet||Yes|
* XMLA Read operations only. Support for XMLA Write operations are coming in early 2020.
Unlimited Power BI Report content viewing is the capability to share a report or dashboard with everyone in an organization. Absent this capability, each user must possess a license to view the dashboard or report.
Paginated Reports is the standard report format in SQL Server Reporting Services and is available in Power BI Premium.
Dataflows provide self-service data prep capabilities for big data in Power BI.
AI Workload lets you use cognitive services and Automated Machine Learning in Power BI.
Multi-model memory management: The total size of models in Azure Analysis Services cannot exceed the memory available in the chosen configuration. In Power BI Premium, models are loaded in memory as required and evicted when not in use providing a more efficient use of memory resources.
Pre-aggregated tables are tables that contain aggregates of other tables. Depending on the size of the base table, queries that can be resolved using the pre-aggregated tables can perform much, much faster than otherwise.
Composite models are individual datasets that support tables with direct query, import, or both – all in a single dataset.
Automated incremental refresh allows datasets to be updated with only changed data rather than refreshing the entire dataset.
Large datasets is the ability to host datasets as large as memory is available. In Power BI Premium, the limit was 10GB but this ceiling has since been removed (this capability is in preview).
Third party application support means applications other than Power BI can be used to query and manage datasets with an XML/A endpoint.
Scale Out enables client query distribution across query replicas, reducing response times for high-concurrency workloads. Scale out is not yet available in Power BI Premium but is planned for mid to late 2020.
Bring Your Own Key gives you the ability to control the encryption keys for your Power BI data at-rest. This is available in preview.
Metadata Translations provide the ability to support many different languages of model names, table names, and column names that are automatically included in reports and dashboards based on a customer’s locale.
Object Level Security provides table and column level security on tabular models.
Perspectives define viewable subsets of a model that provide focused, business-specific, or application-specific viewpoints.
Power BI Premium is, as the name suggests, the premium offering for enterprise and self service BI. Power BI Premium is licensed by capacity and you pay for as much dedicated capacity as you need. Any users can consume Power BI content – only authors require an individual Power BI license. More information on Power BI pricing can be found here.
Azure Analysis Services is available in Developer, Basic and Standard service tiers. Within each tier exists several more instance types with different processing and memory options depending on your data and query processing workload. Detailed pricing information can be found here. You can pause and restart Azure Analysis Services and no costs accrue when it is not running. Any Power BI licenses must be purchased separately.
There are two key takeaways from this blog. First, there are a few capabilities where Azure Analysis Services satisfies requirements not yet generally available (GA) in Power BI Premium. If you require very large datasets or query scale out today, Azure Analysis Services might be your preferred solution. However, these features will soon be matched or exceeded in Power BI Premium.
Second, Power BI Premium is the enterprise platform for modern enterprise cloud BI applications. Analysis Services is the engine powering this platform. While Power BI Premium has (and will continue to have) additional capabilities for performance, scale, and management, the engine itself may satisfy your needs. Both are excellent and the choice between the two options is dependent on your business requirements.