I think the correct link to the article you are linking to is this: http://wildfly.org/news/2014/02/06/GlassFish-to-WildFly-migration/ (The other does not seem to point to the article, which you mention)
The following is an interesting part of the "Database Connection Migration" section that you are referring to:
In WildFly, you have two ways to install the JDBC driver: you deploy it like any other application package or install it as a module. You can always choose a driver deployment, but especially recommend it when you have a clustered environment, since deployments are automatically distributed in server groups. You may have deployment problems if the driver is not JDBC4-compatible. In this case, installing the driver as a module solves these problems. The advantage of the JDBC driver as a module is the ability to create a custom WildFly package for your organization. This way, you can repeat the exact same setup on multiple machines while maintaining the same configuration. This is ideal for a development environment.
So, in this section, the author describes the following advantage:
You may have deployment problems if the driver is not compatible with JDBC4. In this case, installing the driver as a module solves these problems.
The following Wildfly documentation describes this as well:
Any JDBC-compatible driver is automatically recognized and installed on the system by name and version. JDBC JAR is identified using the Java service provider mechanism. Such JARs will contain text with the name META-INF / services / java.sql.Driver that contains the name of the class (s) of drivers that exist in this JAR. If the JDBC JAR driver is not compatible with JDBC 4, you can make it deployable in one of several ways. (...)
Thus, deploying your driver as a module is easier than deploying it like any other application package if it is not compatible with JDBC-4. (Because you will have to modify and rebuild JDBC-4, which is not compatible with it, to deploy it like any other application package)
source share