Connecting to IBM MQ

All network communication with IBM MQ queue managers, including communication related to Managed File Transfer, involves IBM MQ channels. An IBM MQ channel represents one end of a network link. Channels are classified as either message channels or MQI channels.

Managed File Transfer and channels

Managed File Transfer uses MQI channels to connect agents in client mode to their agent queue managers, and to connect command applications (for example, fteCreateTransfer) to their command and coordination queue managers. In the default configuration, these connections are made using a SVRCONN channel called SYSTEM.DEF.SVRCONN, which exists by default on all queue managers. Because of these defaults, we do not need to alter any MQI channels for a basic Managed File Transfer installation.

There are six types of message channel end points, but this topic covers only sender-receiver pairs. See Distributed queuing components for information about other channel combinations.

Required message paths

IBM MQ messages can travel over message channels only, so you must ensure that channels are available for all message paths required by Managed File Transfer. These paths do not need to be direct; messages can travel through intermediate queue managers if required. This topic covers only direct point-to-point communication. See How to get to the remote queue manager for more information about these options. The communication paths used by Managed File Transfer are as follows:

Agent to agent

Any two agents that files are transferred between require bidirectional communication between their associated queue managers. Because this path carries the bulk data, consider making the path as short, fast, or cheap as possible according to your needs.

Agent to coordination

Log messages from the agents that participate in a transfer must be able to reach the coordination queue manager.

Command to agent

Any queue manager that command applications or the IBM MQ Explorer (using the command queue manager) connect to must be able to send messages to the queue managers of the agents that those command applications are used to control. To enable feedback messages to be shown by the commands, use a bidirectional connection.

For more information, see Verifying an IBM MQ installation for the platform, or platforms, that your enterprise use.