JDBC Drivers and datasources can be configured in two different ways within a WildFly Swarm application.
The first is through a deployment, which is, as you might guess, a deployment-time creation of a datasource. Through deployments you can create both JDBC drivers and associated datasources that other deployments can use.
The second is through fraction configuration, which is at a slightly higher level, tied to the container configuration.
While both can result in the same functionality, there is not
any particular difference between them other than how you package
your application and if you want to work with JBoss Modules