Since its launch over 5 years ago, Power BI has empowered self-service business analysts to discover insights quickly and easily without a dependency on IT. Over the years, Power BI evolved into the leading platform for both self-service and IT-managed enterprise BI. This is understood by leading independent research agencies Gartner and Forrester. Power BI is a one-stop shop for all BI users.
With exponential growth in data volumes and complexity, Power BI customers demand enterprise BI solutions that scale to petabytes, are secure, easy to manage, and accessible to all users across the largest organizations. We at Microsoft have a deep heritage in enterprise BI.
Azure Analysis Services (AAS) and SQL Server Analysis Services (SSAS) are based on mature BI engine technology used by countless enterprises. The same technology is also at the heart of Power BI datasets. We have taken great strides to bring AAS capabilities to Power BI. Power BI has inherited a large ecosystem of developers, partners, BI tools, and solutions built up over decades.
Azure Analysis Services capabilities in Power BI Premium
Here are the statuses of notable capabilities that Power BI is gaining from AAS, helping make Power BI a superset of AAS. They are targeted at Power BI Premium, Premium per User and Power BI Embedded.
- XMLA endpoints are now generally available!
In addition to complex semantic modeling, dataset management capabilities and backwards compatibility with Analysis Services tools and processes, XMLA endpoints provide open-platform connectivity for single-version-of-the-truth semantic models. In terms of scenarios covered, XMLA endpoints represents the biggest single capability inherited from Analysis Services. - Large models are now generally available!
The large models feature enables blazing fast user interactivity over vast amounts of data. Dataset size is limited only by the resources on the capacity similarly to AAS models on a server. - Object-level security (OLS) is in public preview!
OLS in Power BI restricts access to tables and columns containing sensitive data such as personally identifiable information (PII). In addition to the data, metadata is protected to help prevent a malicious user from discovering the object exists. - Azure Log Analytics integration is planned for March 2021
Like AAS diagnostic logging, Azure Log Analytics integration with Power BI Premium enables administrators to monitor usage, system health, can be used for diagnosis for performance improvements and auditing purposes. - Backup/restore is planned for April 2021
Backup/restore of datasets in Power BI Premium protects critical data for business continuity planning, disaster recovery, and application lifecycle management (ALM) purposes. It is backwards compatible with backup/restore in AAS. Backup/restore offers a migration path by taking a backup of an AAS model and restoring it to Power BI. - Asynchronous refresh is planned for the 2nd half of 2021 calendar year
The asynchronous refresh REST API for Power BI Premium datasets improves reliability of refresh operations for large datasets with billions of rows for blazing fast interactive analysis. Like asynchronous refresh in AAS, it can be invoked with batched commits. - Query scale out is being evaluated in the context of the Power BI Premium Gen2 architecture, which will redefine the landscape for dataset performance and concurrency. We will provide an update when the evaluation is complete.
- The public preview of external tools in Power BI Desktop means BI professionals with Analysis Services experience can more easily enjoy a range of additional semantic modeling features, DAX query/expression optimization and authoring, and application lifecycle management (ALM) capabilities that were traditionally available only in Analysis Services.
Integration with the Power BI ecosystem for unparalleled capabilities
Customers don’t just want AAS features in Power BI. Enterprise semantic models require native integration with the Power BI ecosystem to leverage the full power of the Power BI platform. A wide range of enterprise features are only available to Power BI datasets.
Here is just a small selection of additional capabilities that happen to be particularly relevant to enterprise semantic models in Power BI.
- Other workloads such as Paginated Reports, Dataflows and AI. Paginated reports allow developers to create and distribute highly formatted, pixel-perfect reports. Dataflows enables self-service data prep for big data. The AI workload empowers business users with Cognitive Services, and AutoML.
- The Power BI performance accelerator for Azure Synapse Analytics is planned for public preview this summer. This capability for massive-scale data warehouses creates materialized views to boost performance by tracking Power BI queries. Performance accelerator is a self-optimizing system that automatically adapts as usage patterns evolve.
- Composite models and aggregations for petabyte-scale interactive analysis. Unlock massive datasets with instant query response times, striking a balance between cost, performance, and data freshness.
- The recently announced Power BI integration with Azure Purview provides metadata discovery, exploration, classification, visibility of data definitions and lineage across the Microsoft data stack.
- Incremental refresh is a critical capability to unlock actionable insights for intelligent decision-making over very large datasets. It is a great example of how Power BI provides a simplified management experience for enterprise BI. Refreshes are faster and more reliable.
- Microsoft Information Protection (MIP) sensitivity labels in Power BI integrate with Microsoft Cloud App Security for data loss prevention. This capability is unique to Power BI among BI vendors. MIP labels are inherited through datasets to other artefacts across the stack including apps, reports, dashboards, Excel PivotTables and export to Excel.
- Bring your own key (BYOK) in Power BI Premium gives enterprises the ability to configure the encryption key used to encrypt their data when it’s stored in the Microsoft cloud for adherence with compliance requirements and enhanced security.
- VNet connectivity for Power BI is planned for March 2021. It will allow Power BI to work seamlessly in a company’s VNet for enhanced security and connect to Azure data sources without the need for a gateway.
- Deployment pipelines provide a no-code/low-code, guided experience for application lifecycle management of Power BI artefacts including datasets. Quickly and easily create repeatable deployment processes to help ensure quality of mission-critical BI systems for thousands of users.
- Shared and certified datasets are sanctioned as the single version of the truth semantic model for your organization, so they are easy for users to find, promote reusability, and standardized decision making.
- The data lineage view in Power BI helps promote trust in data surfaced by Power BI through transparency. It provides an end-to-end 360-degree view from data sources all the way through to dashboards and artefacts in between, and across workspaces. Impact analysis allows change notification for downstream consumers.
- Metadata translations are observed by the Power BI service for datasets in Power BI Premium. Make your Power BI datasets multi-lingual, so names of tables, columns, measures, and other objects are automatically displayed in Power BI reports using the language of the current user.
Power BI Premium Gen2
Perhaps what is generating the most excitement right now for Power BI as the choice for semantic models is the announcement of Power BI Premium Gen2.
Low-cost entry point
Premium Per User (PPU) provides a lower-cost price point to get access to Premium capabilities on the Gen2 platform. This addresses a need we’ve heard time and again from AAS customers. PPU helps remove possibly the biggest barrier to migrations from AAS to Power BI.
Consistent performance
The Gen2 architecture is more insensitive to overall load, temporal spikes, and high concurrency. Gen2 offers significant concurrency and performance gains due to its distributed architecture.
Reduced cost of ownership
With most PaaS services, some management overhead is inevitable. Power BI Premium Gen2 is more like a SaaS service in this respect.
- Gen2 costs are charged based on CPU usage, negating the need to pause servers when not being used.
- Concurrent dataset refreshes on the same virtual capacity are automatically distributed across physical machines, negating the need to manage dataset contention.
- Especially when consolidating capacities to larger SKUs, the need for scale up/down can be reduced.
Why converge enterprise and self-service BI?
Consolidation of artefacts in Power BI can result in simplified discovery and management due to co-location. There is no need to “bridge the gap” between multiple products. Central IT teams can more easily harvest self-service artefacts that have become popular and are causing a management burden for the business. Such artefacts are taken over by IT and operationalized for mission-critical decision making based on governed data, alignment with corporate standards, and lineage transparency. Simplifying this workflow by sharing a common platform promotes collaboration between the business and IT.
Customer stories
Here are just a few examples of customers who are benefiting the Power BI platform for enterprise semantic models.
 | Hexagon is a global leader in sensor, software, and autonomous solutions. Hexagon’s Safety & Infrastructure division provides software to improve the performance, efficiency, and resilience of vital services for public safety, transportation, utilities, and more. The semantic models were migrated from SSAS (on-premises) to Power BI datasets in the cloud to ensure a single source of truth for every user. For more information, read the Hexagon customer story. |
 | Arla Foods is a Danish-Swedish multinational cooperative and the fifth-largest dairy company in the world. Arla Foods uses Power BI datasets as rich semantic models spanning multiple subject areas, empowering business users to interactively create beautiful reports. Easy access to curated data through semantic models promotes consistent decision making in alignment with Arla Foods’s strategies and policies. For more information, read the Arla Foods customer story. |
 | The Forsyth County School District is the seventh-largest district in the state of Georgia with 40 schools and centers, over 8,000 full-time and part-time employees, and more than 51,000 students. Semantic models in Power BI provide single source of truth data for numerous functions including identification of students needing intervention to improve academic success. For more information, read the Forsyth County Schools customer story. |
 | The Danish Agriculture & Food Council’s subsidiary, SEGES is the leading agricultural advisor in Denmark. A semantic model in Power BI represents a single source of truth for reporting and data visualization using Power BI Embedded in a web portal. For more information, read the SEGES customer story. |
What’s next for AAS?
As we make Power BI a superset of AAS, we understand it’s not always easy for existing AAS customers to transition to Power BI Premium. For example, Azure billing is preferred by some AAS customers. There are manual steps required for migration. The different licensing models and costs of AAS and Power BI Premium can make it difficult. We are working on a plan to address this and will share it in the coming months. Please stay tuned. I guarantee you won’t want to miss it!
Closing thoughts
To be clear and transparent, we will continue to invest in Power BI Premium for enterprise semantic models. The full set of Power BI workloads, features and capabilities represent a modern, cloud-born BI platform that goes far beyond comparable functionality available on premises. This is appealing to a large proportion of BI customers moving to the cloud. We’d like to express sincere appreciation to all our enterprise BI customers, as we look forward to continuing our journey together to make Power BI a superset of Analysis Services. Here is a visual representation focusing on dataset related capabilities.
