Once we have the logical data models, we are ready to create physical data models. This is the stage where you convert the logical data models into a database design for a specific RDBMS (short for Rational Database Management System), such as MySQL or PostgreSQL, or a NoSQL database, such as MongoDB. An important consideration during physical data modeling is performance. Most of the database systems support tools such as indexing, clustering, partitioning, and data compression to archive performance improvement without compromising the logical data model design. We will also discuss indexing in this book. But we will not discuss clustering, portioning, or data compression. They are beyond the scope of this book. 

The output of this stage are the physical data models, and since we're using MySQL (5.7.21) as our database, we will use MySQL Workbench (https://www.mysql.com/products/workbench) for the physical data modeling. One of the nice features that MySQL Workbench supports is the ability to export physical data models as SQL that we can use to set up the database.

The following is an activity diagram that shows different data modeling stages and the deliverable of each stage. The round-cornered rectangles represent actions, such as creating Conceptual Data Model, while the square-cornered rectangles represent the object flows between actions, for example, the Application Requirements node is the input of creating Conceptual Data Model: