IBM Data Virtualization accesses files in shared directories (network file systems) by using a remote connector. This connector is a Java-based application that is installed and run on a server within the remote data source's network. It acts as a secure gateway, allowing the Data Virtualization service, which runs on the Cloud Pak for Data cluster, to connect to and query data from file systems that are not directly accessible, such as those behind a firewall. This architecture enables the virtualization of data without moving it.

A. Adding a connection in the UI is part of the workflow, but it's a user action that relies on the pre-installed remote connector to function.

B. The service does not automatically scan networks for security and performance reasons; connections must be explicitly configured by an administrator.

C. The remote connector is a general-purpose agent and is not limited to the FTP protocol; it supports various types of remote data sources.

1. IBM Cloud Pak for Data 4.6 Documentation

"Adding data from files in Data Virtualization": This document states

"To virtualize data from files in a network file system

you must first install and configure a remote connector." This directly confirms that the remote connector is the required mechanism.

2. IBM Cloud Pak for Data 4.6 Documentation

"Installing the remote connector": This section describes the remote connector's purpose: "The Data Virtualization remote connector is a Java-based application that you can deploy on Windows or Linux®. The remote connector connects to data sources that are on-premises or in other clouds." This explains the role and nature of the connector used for accessing remote sources like shared directories.