Choosing Azure Analysis Services for large datasets
I often get questions about use cases for Azure Analysis Services. So I wanted to do a quick recap of the points I commonly discuss in these conversations, and have them readily available for the next conversation… As a bonus, I get to share them with a larger audience!
The Tabular semantic model was introduced to a large number of Power BI users, and they take advantage of the Analysis Services technology starting with the free Power BI Desktop. Indeed, the same engine runs Power BI, Power Pivot for Excel, SQL Server Analysis Services, and Azure Analysis Services. However, when dealing with a large dataset, you want to use the engine on a separate backend infrastructure. By large I mean databases with one or more tables over 200 million rows. The database size is just a consequence of the table size. Many of this article’s considerations do not apply to large models that just have many columns and/or many unique values in smaller tables.
Tags: performance, tabular, azure