Serving Layer in Azure

Categories of Data stores

1. Key/ value DBs : Holds a single serialized object for each key value
2. Document : Key / value databases with document as value
3. Column-family: Key / value databases that structure data storage into collections of related columns
4. Graph: Store information as a collection of objects and relationships
5. Telemetry and time series: Append only collection of objects

Understanding the STAR schema

To understand some star schema concepts described in this article, it's important to know two terms: normalization and denormalization. Normalization is the term used to describe data that's stored in a way that reduces repetitive data. Consider a table of products that has a unique key value column, like the product key, and additional columns describing product characteristics, including product name, category, color, and size. A sales table is considered normalized when it stores only keys, like the product key. In the following image, notice that only the ProductKey column records the product.
If, however, the sales table stores product details beyond the key, it's considered denormalized. In the following image, notice that the ProductKey and other product-related columns record the product.

When you source data from an export file or data extract, it's likely that it represents a denormalized set of data.

For slowly changing dimensions, Role Playing dimensions, Junk dimensions read this https://docs.microsoft.com/en-us/power-bi/guidance/star-schema